menekis/OCRmyPDF
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Regroup three merge steps into a single step

Browse Source 
All take the same inputs and deliver similar outputs, so it makes sense.
This commit is contained in:
James R. Barlow
2018-04-06 01:07:02 -07:00
parent 4a341c9034
commit 709c01c7a1
2 changed files with 435 additions and 541 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/pipeline.svg
+404 -484
View File
@@ -4,550 +4,470 @@
<!-- Generated by graphviz version 2.40.1 (20161225.0304)
-->
<!-- Title: Pipeline: Pages: 1 -->
<svg width="1952pt" height="1277pt"
viewBox="0.00 0.00 1952.00 1277.18" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 1273.1825)">
<svg width="1932pt" height="1277pt"
viewBox="0.00 0.00 1932.00 1277.21" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink">
<g id="graph0" class="graph" transform="scale(1 1) rotate(0) translate(4 1273.2069)">
<title>Pipeline:</title>
<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-1273.1825 1948,-1273.1825 1948,4 -4,4"/>
<polygon fill="#ffffff" stroke="transparent" points="-4,4 -4,-1273.2069 1928,-1273.2069 1928,4 -4,4"/>
<g id="clust1" class="cluster">
<title>clustertasks</title>
<polygon fill="none" stroke="#000000" points="8,-8 8,-1261.1825 1936,-1261.1825 1936,-8 8,-8"/>
<text text-anchor="middle" x="972" y="-1233.1825" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
<polygon fill="none" stroke="#000000" points="8,-8 8,-1261.2069 1916,-1261.2069 1916,-8 8,-8"/>
<text text-anchor="middle" x="962" y="-1233.2069" font-family="Times,serif" font-size="30.00" fill="#ff3232">Pipeline:</text>
</g>
<!-- t0 -->
<g id="node1" class="node">
<title>t0</title>
<polygon fill="#efa03b" stroke="#000000" points="876.7053,-1215.1825 667.2947,-1215.1825 663.2947,-1211.1825 663.2947,-1179.1825 872.7053,-1179.1825 876.7053,-1183.1825 876.7053,-1215.1825"/>
<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 663.2947,-1211.1825 "/>
<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 872.7053,-1179.1825 "/>
<polyline fill="none" stroke="#000000" points="872.7053,-1211.1825 876.7053,-1215.1825 "/>
<text text-anchor="middle" x="770" y="-1191.1825" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.triage</text>
<polygon fill="#efa03b" stroke="#000000" points="1810.7053,-1215.2069 1601.2947,-1215.2069 1597.2947,-1211.2069 1597.2947,-1179.2069 1806.7053,-1179.2069 1810.7053,-1183.2069 1810.7053,-1215.2069"/>
<polyline fill="none" stroke="#000000" points="1806.7053,-1211.2069 1597.2947,-1211.2069 "/>
<polyline fill="none" stroke="#000000" points="1806.7053,-1211.2069 1806.7053,-1179.2069 "/>
<polyline fill="none" stroke="#000000" points="1806.7053,-1211.2069 1810.7053,-1215.2069 "/>
<text text-anchor="middle" x="1704" y="-1191.2069" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.triage</text>
</g>
<!-- t1 -->
<g id="node2" class="node">
<title>t1</title>
<polygon fill="#efa03b" stroke="#000000" points="941.0433,-1157.1825 602.9567,-1157.1825 598.9567,-1153.1825 598.9567,-1121.1825 937.0433,-1121.1825 941.0433,-1125.1825 941.0433,-1157.1825"/>
<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 598.9567,-1153.1825 "/>
<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 937.0433,-1121.1825 "/>
<polyline fill="none" stroke="#000000" points="937.0433,-1153.1825 941.0433,-1157.1825 "/>
<text text-anchor="middle" x="770" y="-1133.1825" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.repair_and_parse_pdf</text>
<polygon fill="#efa03b" stroke="#000000" points="1875.0433,-1157.2069 1536.9567,-1157.2069 1532.9567,-1153.2069 1532.9567,-1121.2069 1871.0433,-1121.2069 1875.0433,-1125.2069 1875.0433,-1157.2069"/>
<polyline fill="none" stroke="#000000" points="1871.0433,-1153.2069 1532.9567,-1153.2069 "/>
<polyline fill="none" stroke="#000000" points="1871.0433,-1153.2069 1871.0433,-1121.2069 "/>
<polyline fill="none" stroke="#000000" points="1871.0433,-1153.2069 1875.0433,-1157.2069 "/>
<text text-anchor="middle" x="1704" y="-1133.2069" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.repair_and_parse_pdf</text>
</g>
<!-- t0&#45;&gt;t1 -->
<g id="edge1" class="edge">
<title>t0&#45;&gt;t1</title>
<path fill="none" stroke="#0044a0" d="M770,-1179.1488C770,-1175.4743 770,-1171.5242 770,-1167.5888"/>
<polygon fill="#0044a0" stroke="#0044a0" points="773.5001,-1167.4047 770,-1157.4047 766.5001,-1167.4047 773.5001,-1167.4047"/>
<path fill="none" stroke="#0044a0" d="M1704,-1179.1733C1704,-1175.4987 1704,-1171.5487 1704,-1167.6132"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1707.5001,-1167.4291 1704,-1157.4291 1700.5001,-1167.4292 1707.5001,-1167.4291"/>
</g>
<!-- t2 -->
<g id="node3" class="node">
<title>t2</title>
<polygon fill="#efa03b" stroke="#000000" points="1043.9447,-1078.2585 787,-1095.158 530.0553,-1078.2585 530.2951,-1050.9147 1043.7049,-1050.9147 1043.9447,-1078.2585"/>
<polygon fill="none" stroke="#000000" points="1047.9814,-1081.9991 787,-1099.164 526.0186,-1081.9991 526.3263,-1046.9173 1047.6737,-1046.9173 1047.9814,-1081.9991"/>
<text text-anchor="middle" x="787" y="-1064.7009" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.pre_split_pages</text>
<polygon fill="#efa03b" stroke="#000000" points="1656.9447,-1078.283 1400,-1095.1824 1143.0553,-1078.283 1143.2951,-1050.9391 1656.7049,-1050.9391 1656.9447,-1078.283"/>
<polygon fill="none" stroke="#000000" points="1660.9814,-1082.0236 1400,-1099.1885 1139.0186,-1082.0236 1139.3263,-1046.9417 1660.6737,-1046.9417 1660.9814,-1082.0236"/>
<text text-anchor="middle" x="1400" y="-1064.7253" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.pre_split_pages</text>
</g>
<!-- t1&#45;&gt;t2 -->
<g id="edge2" class="edge">
<title>t1&#45;&gt;t2</title>
<path fill="none" stroke="#0044a0" d="M774.5557,-1120.8306C775.4962,-1117.0421 776.5194,-1112.92 777.5618,-1108.7212"/>
<polygon fill="#0044a0" stroke="#0044a0" points="780.986,-1109.4542 779.9984,-1098.9055 774.1922,-1107.7677 780.986,-1109.4542"/>
<path fill="none" stroke="#0044a0" d="M1623.7299,-1121.1246C1588.2849,-1113.14 1546.0878,-1103.6343 1507.9425,-1095.0414"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1508.3459,-1091.5446 1497.8212,-1092.7613 1506.8075,-1098.3735 1508.3459,-1091.5446"/>
</g>
<!-- t20 -->
<g id="node22" class="node">
<title>t20</title>
<polygon fill="#efa03b" stroke="#000000" points="649.5156,-1020.2193 286.4844,-1020.2193 282.4844,-1016.2193 282.4844,-984.2193 645.5156,-984.2193 649.5156,-988.2193 649.5156,-1020.2193"/>
<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 282.4844,-1016.2193 "/>
<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 645.5156,-984.2193 "/>
<polyline fill="none" stroke="#000000" points="645.5156,-1016.2193 649.5156,-1020.2193 "/>
<text text-anchor="middle" x="466" y="-996.2193" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.generate_postscript_stub</text>
<polygon fill="#efa03b" stroke="#000000" points="1887.5156,-1020.2438 1524.4844,-1020.2438 1520.4844,-1016.2438 1520.4844,-984.2438 1883.5156,-984.2438 1887.5156,-988.2438 1887.5156,-1020.2438"/>
<polyline fill="none" stroke="#000000" points="1883.5156,-1016.2438 1520.4844,-1016.2438 "/>
<polyline fill="none" stroke="#000000" points="1883.5156,-1016.2438 1883.5156,-984.2438 "/>
<polyline fill="none" stroke="#000000" points="1883.5156,-1016.2438 1887.5156,-1020.2438 "/>
<text text-anchor="middle" x="1704" y="-996.2438" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.generate_postscript_stub</text>
</g>
<!-- t1&#45;&gt;t20 -->
<g id="edge31" class="edge">
<title>t1&#45;&gt;t20</title>
<path fill="none" stroke="#0044a0" d="M598.785,-1123.5822C563.105,-1117.7214 531.8209,-1109.8012 517,-1099.1825 493.9382,-1082.6593 480.4369,-1052.4741 473.1842,-1030.1561"/>
<polygon fill="#0044a0" stroke="#0044a0" points="476.5199,-1029.095 470.2797,-1020.5329 469.8185,-1031.1177 476.5199,-1029.095"/>
</g>
<!-- t24 -->
<g id="node24" class="node">
<title>t24</title>
<polygon fill="#efa03b" stroke="#000000" points="20.4213,-156.4247 242,-129.112 463.5787,-156.4247 463.3719,-200.6177 20.6281,-200.6177 20.4213,-156.4247"/>
<polygon fill="none" stroke="#000000" points="16.4083,-152.8864 242,-125.079 467.5917,-152.8864 467.3496,-204.6205 16.6504,-204.6205 16.4083,-152.8864"/>
<text text-anchor="middle" x="242" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_mupdf</text>
</g>
<!-- t1&#45;&gt;t24 -->
<g id="edge41" class="edge">
<title>t1&#45;&gt;t24</title>
<path fill="none" stroke="#0044a0" d="M598.6617,-1124.8567C376.4578,-1103.9215 17,-1060.9746 17,-1002.2193 17,-1002.2193 17,-1002.2193 17,-335.7123 17,-276.266 68.2524,-235.5065 121.8346,-209.1377"/>
<polygon fill="#0044a0" stroke="#0044a0" points="123.6227,-212.1629 131.1479,-204.7049 120.6143,-205.8423 123.6227,-212.1629"/>
</g>
<!-- t23 -->
<g id="node25" class="node">
<title>t23</title>
<polygon fill="#efa03b" stroke="#000000" points="1006.4902,-156.4239 1218,-129.1092 1429.5098,-156.4239 1429.3124,-200.6199 1006.6876,-200.6199 1006.4902,-156.4239"/>
<polygon fill="none" stroke="#000000" points="1002.4742,-152.9063 1218,-125.073 1433.5258,-152.9063 1433.2948,-204.623 1002.7052,-204.623 1002.4742,-152.9063"/>
<text text-anchor="middle" x="1218" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_qpdf</text>
</g>
<!-- t1&#45;&gt;t23 -->
<g id="edge46" class="edge">
<title>t1&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M941.3419,-1136.2202C1263.4143,-1128.3561 1927,-1100.2573 1927,-1002.2193 1927,-1002.2193 1927,-1002.2193 1927,-335.7123 1927,-234.7734 1651.3069,-194.7548 1443.7522,-178.9295"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1443.8443,-175.4267 1433.6111,-178.1712 1443.3223,-182.4072 1443.8443,-175.4267"/>
</g>
<!-- t3 -->
<g id="node4" class="node">
<title>t3</title>
<polygon fill="#efa03b" stroke="#000000" points="1055.3851,-1020.2193 808.6149,-1020.2193 804.6149,-1016.2193 804.6149,-984.2193 1051.3851,-984.2193 1055.3851,-988.2193 1055.3851,-1020.2193"/>
<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 804.6149,-1016.2193 "/>
<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 1051.3851,-984.2193 "/>
<polyline fill="none" stroke="#000000" points="1051.3851,-1016.2193 1055.3851,-1020.2193 "/>
<text text-anchor="middle" x="930" y="-996.2193" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.split_page</text>
</g>
<!-- t2&#45;&gt;t3 -->
<g id="edge3" class="edge">
<title>t2&#45;&gt;t3</title>
<path fill="none" stroke="#0044a0" d="M837.0145,-1046.7494C851.946,-1039.5987 868.2558,-1031.7881 882.9879,-1024.733"/>
<polygon fill="#0044a0" stroke="#0044a0" points="884.9035,-1027.6964 892.4109,-1020.2204 881.88,-1021.383 884.9035,-1027.6964"/>
</g>
<!-- t4 -->
<g id="node5" class="node">
<title>t4</title>
<polygon fill="#efa03b" stroke="#000000" points="1194.5194,-941.2939 964,-958.1945 733.4806,-941.2939 733.6958,-913.948 1194.3042,-913.948 1194.5194,-941.2939"/>
<polygon fill="none" stroke="#000000" points="1198.5491,-945.0069 964,-962.203 729.4509,-945.0069 729.7267,-909.9503 1198.2733,-909.9503 1198.5491,-945.0069"/>
<text text-anchor="middle" x="964" y="-927.7357" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_or_skip</text>
</g>
<!-- t3&#45;&gt;t4 -->
<g id="edge4" class="edge">
<title>t3&#45;&gt;t4</title>
<path fill="none" stroke="#0044a0" d="M939.1114,-983.867C941.197,-979.6659 943.4863,-975.0549 945.8053,-970.3839"/>
<polygon fill="#0044a0" stroke="#0044a0" points="948.9439,-971.9327 950.2559,-961.4194 942.6741,-968.8199 948.9439,-971.9327"/>
</g>
<!-- t5 -->
<g id="node6" class="node">
<title>t5</title>
<polygon fill="#efa03b" stroke="#000000" points="1119.3166,-883.252 812.6834,-883.252 808.6834,-879.252 808.6834,-847.252 1115.3166,-847.252 1119.3166,-851.252 1119.3166,-883.252"/>
<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 808.6834,-879.252 "/>
<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 1115.3166,-847.252 "/>
<polyline fill="none" stroke="#000000" points="1115.3166,-879.252 1119.3166,-883.252 "/>
<text text-anchor="middle" x="964" y="-859.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_preview</text>
</g>
<!-- t4&#45;&gt;t5 -->
<g id="edge5" class="edge">
<title>t4&#45;&gt;t5</title>
<path fill="none" stroke="#0044a0" d="M964,-909.7834C964,-904.5184 964,-898.8955 964,-893.4799"/>
<polygon fill="#0044a0" stroke="#0044a0" points="967.5001,-893.4695 964,-883.4695 960.5001,-893.4695 967.5001,-893.4695"/>
</g>
<!-- t6 -->
<g id="node7" class="node">
<title>t6</title>
<polygon fill="#efa03b" stroke="#000000" points="1269.0824,-825.252 1010.9176,-825.252 1006.9176,-821.252 1006.9176,-789.252 1265.0824,-789.252 1269.0824,-793.252 1269.0824,-825.252"/>
<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1006.9176,-821.252 "/>
<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1265.0824,-789.252 "/>
<polyline fill="none" stroke="#000000" points="1265.0824,-821.252 1269.0824,-825.252 "/>
<text text-anchor="middle" x="1138" y="-801.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.orient_page</text>
</g>
<!-- t4&#45;&gt;t6 -->
<g id="edge7" class="edge">
<title>t4&#45;&gt;t6</title>
<path fill="none" stroke="#0044a0" d="M1072.6871,-909.8532C1097.768,-902.2558 1119.5887,-893.2079 1128,-883.252 1138.9559,-870.2843 1141.4763,-851.3696 1141.2475,-835.6545"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1144.7291,-835.2052 1140.7199,-825.3983 1137.7384,-835.5649 1144.7291,-835.2052"/>
</g>
<!-- t5&#45;&gt;t6 -->
<g id="edge6" class="edge">
<title>t5&#45;&gt;t6</title>
<path fill="none" stroke="#0044a0" d="M1018.1009,-847.2184C1035.9411,-841.2716 1055.9459,-834.6034 1074.4267,-828.4431"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1075.5526,-831.7572 1083.9326,-825.2745 1073.3389,-825.1165 1075.5526,-831.7572"/>
</g>
<!-- t7 -->
<g id="node8" class="node">
<title>t7</title>
<polygon fill="#efa03b" stroke="#000000" points="917.7325,-767.252 542.2675,-767.252 538.2675,-763.252 538.2675,-731.252 913.7325,-731.252 917.7325,-735.252 917.7325,-767.252"/>
<polyline fill="none" stroke="#000000" points="913.7325,-763.252 538.2675,-763.252 "/>
<polyline fill="none" stroke="#000000" points="913.7325,-763.252 913.7325,-731.252 "/>
<polyline fill="none" stroke="#000000" points="913.7325,-763.252 917.7325,-767.252 "/>
<text text-anchor="middle" x="728" y="-743.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_with_ghostscript</text>
</g>
<!-- t6&#45;&gt;t7 -->
<g id="edge8" class="edge">
<title>t6&#45;&gt;t7</title>
<path fill="none" stroke="#0044a0" d="M1010.5209,-789.2184C964.4341,-782.6988 912.2171,-775.312 865.3222,-768.6781"/>
<polygon fill="#0044a0" stroke="#0044a0" points="865.792,-765.2098 855.4003,-767.2745 864.8114,-772.1408 865.792,-765.2098"/>
</g>
<!-- t17 -->
<g id="node15" class="node">
<title>t17</title>
<polygon fill="#ff69b4" stroke="#000000" points="1224.2696,-447.252 843.7304,-447.252 839.7304,-443.252 839.7304,-411.252 1220.2696,-411.252 1224.2696,-415.252 1224.2696,-447.252"/>
<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 839.7304,-443.252 "/>
<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 1220.2696,-411.252 "/>
<polyline fill="none" stroke="#000000" points="1220.2696,-443.252 1224.2696,-447.252 "/>
<text text-anchor="middle" x="1032" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_textonly_pdf</text>
</g>
<!-- t6&#45;&gt;t17 -->
<g id="edge16" class="edge">
<title>t6&#45;&gt;t17</title>
<path fill="none" stroke="#0044a0" d="M1158.4397,-788.9107C1179.7261,-767.6923 1210,-730.5987 1210,-691.252 1210,-691.252 1210,-691.252 1210,-575.252 1210,-532.3545 1227.3937,-512.2638 1200,-479.252 1190.5284,-467.8379 1163.9108,-457.8666 1134.5783,-449.8944"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1135.1079,-446.4151 1124.5475,-447.2681 1133.3349,-453.1869 1135.1079,-446.4151"/>
</g>
<!-- t14 -->
<g id="node17" class="node">
<title>t14</title>
<polygon fill="#00cc66" stroke="#000000" points="1565,-457.252 1242.4141,-429.252 1565,-401.252 1887.5859,-429.252 1565,-457.252"/>
<text text-anchor="middle" x="1565" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_image_layer</text>
</g>
<!-- t6&#45;&gt;t14 -->
<g id="edge22" class="edge">
<title>t6&#45;&gt;t14</title>
<path fill="none" stroke="#0044a0" d="M1211.6459,-789.1984C1287.5889,-768.3738 1395,-731.7602 1395,-691.252 1395,-691.252 1395,-691.252 1395,-575.252 1395,-519.4485 1448.3697,-480.3436 1495.3026,-456.7437"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1497.1229,-459.7505 1504.5857,-452.2301 1494.0619,-453.4552 1497.1229,-459.7505"/>
</g>
<!-- t21 -->
<g id="node20" class="node">
<title>t21</title>
<polygon fill="#efa03b" stroke="#000000" points="1414.7716,-270.1725 1169.2284,-270.1725 1165.2284,-266.1725 1165.2284,-234.1725 1410.7716,-234.1725 1414.7716,-238.1725 1414.7716,-270.1725"/>
<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1165.2284,-266.1725 "/>
<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1410.7716,-234.1725 "/>
<polyline fill="none" stroke="#000000" points="1410.7716,-266.1725 1414.7716,-270.1725 "/>
<text text-anchor="middle" x="1290" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.skip_page</text>
</g>
<!-- t6&#45;&gt;t21 -->
<g id="edge28" class="edge">
<title>t6&#45;&gt;t21</title>
<path fill="none" stroke="#0044a0" d="M1269.322,-800.6326C1488.9576,-787.7851 1907,-754.9633 1907,-691.252 1907,-691.252 1907,-691.252 1907,-429.252 1907,-329.7211 1606.6136,-282.8737 1425.0278,-263.647"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1425.1684,-260.1428 1414.8592,-262.586 1424.4419,-267.105 1425.1684,-260.1428"/>
</g>
<!-- t19 -->
<g id="node21" class="node">
<title>t19</title>
<polygon fill="#66ccff" stroke="#000000" points="465.6739,-447.252 60.3261,-447.252 56.3261,-443.252 56.3261,-411.252 461.6739,-411.252 465.6739,-415.252 465.6739,-447.252"/>
<polyline fill="none" stroke="#000000" points="461.6739,-443.252 56.3261,-443.252 "/>
<polyline fill="none" stroke="#000000" points="461.6739,-443.252 461.6739,-411.252 "/>
<polyline fill="none" stroke="#000000" points="461.6739,-443.252 465.6739,-447.252 "/>
<text text-anchor="middle" x="261" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_and_render_pdf</text>
</g>
<!-- t6&#45;&gt;t19 -->
<g id="edge30" class="edge">
<title>t6&#45;&gt;t19</title>
<path fill="none" stroke="#0044a0" d="M1006.6775,-802.4744C883.1032,-797.0687 693.0454,-786.4114 529,-767.252 327.1018,-743.6717 78,-894.5226 78,-691.252 78,-691.252 78,-691.252 78,-575.252 78,-532.3545 60.4757,-512.1551 88,-479.252 97.684,-467.6756 124.6714,-457.7191 154.5327,-449.8074"/>
<polygon fill="#0044a0" stroke="#0044a0" points="155.4872,-453.176 164.3106,-447.311 153.7556,-446.3936 155.4872,-453.176"/>
</g>
<!-- t8 -->
<g id="node9" class="node">
<title>t8</title>
<polygon fill="#efa03b" stroke="#000000" points="942.0801,-709.252 517.9199,-709.252 513.9199,-705.252 513.9199,-673.252 938.0801,-673.252 942.0801,-677.252 942.0801,-709.252"/>
<polyline fill="none" stroke="#000000" points="938.0801,-705.252 513.9199,-705.252 "/>
<polyline fill="none" stroke="#000000" points="938.0801,-705.252 938.0801,-673.252 "/>
<polyline fill="none" stroke="#000000" points="938.0801,-705.252 942.0801,-709.252 "/>
<text text-anchor="middle" x="728" y="-685.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_remove_background</text>
</g>
<!-- t7&#45;&gt;t8 -->
<g id="edge9" class="edge">
<title>t7&#45;&gt;t8</title>
<path fill="none" stroke="#0044a0" d="M728,-731.2184C728,-727.5438 728,-723.5938 728,-719.6583"/>
<polygon fill="#0044a0" stroke="#0044a0" points="731.5001,-719.4742 728,-709.4743 724.5001,-719.4743 731.5001,-719.4742"/>
</g>
<!-- t13 -->
<g id="node16" class="node">
<title>t13</title>
<polygon fill="#efa03b" stroke="#000000" points="481,-535.252 97.1874,-507.252 481,-479.252 864.8126,-507.252 481,-535.252"/>
<text text-anchor="middle" x="481" y="-501.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_visible_page_image</text>
</g>
<!-- t7&#45;&gt;t13 -->
<g id="edge20" class="edge">
<title>t7&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M592.7552,-731.2177C555.8805,-724.9263 521.5464,-717.3855 505,-709.252 470.5499,-692.3178 456.557,-686.3456 441,-651.252 425.2764,-615.7827 441.5685,-572.6191 457.8398,-542.8513"/>
<polygon fill="#0044a0" stroke="#0044a0" points="460.9365,-544.4864 462.8533,-534.0664 454.8568,-541.0167 460.9365,-544.4864"/>
</g>
<!-- t9 -->
<g id="node10" class="node">
<title>t9</title>
<polygon fill="#efa03b" stroke="#000000" points="773.6759,-651.252 454.3241,-651.252 450.3241,-647.252 450.3241,-615.252 769.6759,-615.252 773.6759,-619.252 773.6759,-651.252"/>
<polyline fill="none" stroke="#000000" points="769.6759,-647.252 450.3241,-647.252 "/>
<polyline fill="none" stroke="#000000" points="769.6759,-647.252 769.6759,-615.252 "/>
<polyline fill="none" stroke="#000000" points="769.6759,-647.252 773.6759,-651.252 "/>
<text text-anchor="middle" x="612" y="-627.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_deskew</text>
</g>
<!-- t8&#45;&gt;t9 -->
<g id="edge10" class="edge">
<title>t8&#45;&gt;t9</title>
<path fill="none" stroke="#0044a0" d="M691.9327,-673.2184C680.9122,-667.7081 668.6526,-661.5783 657.1104,-655.8072"/>
<polygon fill="#0044a0" stroke="#0044a0" points="658.5545,-652.6162 648.045,-651.2745 655.424,-658.8772 658.5545,-652.6162"/>
</g>
<!-- t8&#45;&gt;t13 -->
<g id="edge19" class="edge">
<title>t8&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M763.3121,-673.1037C771.2083,-667.2594 778.5531,-659.9847 783,-651.252 790.2604,-636.9942 792.214,-628.3326 783,-615.252 759.3602,-581.6921 662.7714,-550.9641 585.1057,-530.9385"/>
<polygon fill="#0044a0" stroke="#0044a0" points="585.806,-527.5052 575.2515,-528.4283 584.078,-534.2886 585.806,-527.5052"/>
</g>
<!-- t10 -->
<g id="node11" class="node">
<title>t10</title>
<polygon fill="#efa03b" stroke="#000000" points="1117.8828,-593.252 816.1172,-593.252 812.1172,-589.252 812.1172,-557.252 1113.8828,-557.252 1117.8828,-561.252 1117.8828,-593.252"/>
<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 812.1172,-589.252 "/>
<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 1113.8828,-557.252 "/>
<polyline fill="none" stroke="#000000" points="1113.8828,-589.252 1117.8828,-593.252 "/>
<text text-anchor="middle" x="965" y="-569.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_clean</text>
</g>
<!-- t9&#45;&gt;t10 -->
<g id="edge11" class="edge">
<title>t9&#45;&gt;t10</title>
<path fill="none" stroke="#0044a0" d="M721.7564,-615.2184C760.9379,-608.7806 805.2658,-601.4973 845.2464,-594.9283"/>
<polygon fill="#0044a0" stroke="#0044a0" points="846.0113,-598.3496 855.3115,-593.2745 844.8763,-591.4422 846.0113,-598.3496"/>
</g>
<!-- t9&#45;&gt;t13 -->
<g id="edge18" class="edge">
<title>t9&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M593.2309,-615.1993C573.0988,-595.8356 540.6771,-564.6514 515.8453,-540.7673"/>
<polygon fill="#0044a0" stroke="#0044a0" points="518.0381,-538.0203 508.4046,-533.6106 513.1856,-543.0654 518.0381,-538.0203"/>
</g>
<!-- t11 -->
<g id="node12" class="node">
<title>t11</title>
<polygon fill="#efa03b" stroke="#000000" points="1191.1094,-525.252 886.8906,-525.252 882.8906,-521.252 882.8906,-489.252 1187.1094,-489.252 1191.1094,-493.252 1191.1094,-525.252"/>
<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 882.8906,-521.252 "/>
<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 1187.1094,-489.252 "/>
<polyline fill="none" stroke="#000000" points="1187.1094,-521.252 1191.1094,-525.252 "/>
<text text-anchor="middle" x="1037" y="-501.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_ocr_image</text>
</g>
<!-- t10&#45;&gt;t11 -->
<g id="edge12" class="edge">
<title>t10&#45;&gt;t11</title>
<path fill="none" stroke="#0044a0" d="M984.2947,-557.0293C992.2846,-549.4832 1001.6851,-540.6049 1010.318,-532.4517"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1012.7851,-534.9359 1017.6521,-525.5251 1007.9787,-529.8468 1012.7851,-534.9359"/>
</g>
<!-- t10&#45;&gt;t13 -->
<g id="edge17" class="edge">
<title>t10&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M836.5676,-557.2078C770.772,-547.9638 690.4915,-536.6847 622.8731,-527.1846"/>
<polygon fill="#0044a0" stroke="#0044a0" points="622.9531,-523.6616 612.5634,-525.7361 621.9792,-530.5935 622.9531,-523.6616"/>
</g>
<!-- t12 -->
<g id="node13" class="node">
<title>t12</title>
<polygon fill="#00cc66" stroke="#000000" points="822.0996,-447.252 507.9004,-447.252 503.9004,-443.252 503.9004,-411.252 818.0996,-411.252 822.0996,-415.252 822.0996,-447.252"/>
<polyline fill="none" stroke="#000000" points="818.0996,-443.252 503.9004,-443.252 "/>
<polyline fill="none" stroke="#000000" points="818.0996,-443.252 818.0996,-411.252 "/>
<polyline fill="none" stroke="#000000" points="818.0996,-443.252 822.0996,-447.252 "/>
<text text-anchor="middle" x="663" y="-423.252" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_hocr</text>
</g>
<!-- t11&#45;&gt;t12 -->
<g id="edge13" class="edge">
<title>t11&#45;&gt;t12</title>
<path fill="none" stroke="#0044a0" d="M950.2568,-489.1612C893.2425,-477.2705 818.5749,-461.6981 759.6569,-449.4104"/>
<polygon fill="#0044a0" stroke="#0044a0" points="760.1,-445.9276 749.5961,-447.3122 758.6708,-452.7802 760.1,-445.9276"/>
</g>
<!-- t11&#45;&gt;t17 -->
<g id="edge15" class="edge">
<title>t11&#45;&gt;t17</title>
<path fill="none" stroke="#0044a0" d="M1035.8403,-489.1612C1035.2388,-479.7779 1034.4904,-468.1019 1033.8144,-457.5566"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1037.3033,-457.2705 1033.1707,-447.5149 1030.3177,-457.7184 1037.3033,-457.2705"/>
</g>
<!-- t15 -->
<g id="node14" class="node">
<title>t15</title>
<polygon fill="#00cc66" stroke="#000000" points="1229.0433,-353.7123 920.9567,-353.7123 916.9567,-349.7123 916.9567,-317.7123 1225.0433,-317.7123 1229.0433,-321.7123 1229.0433,-353.7123"/>
<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 916.9567,-349.7123 "/>
<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 1225.0433,-317.7123 "/>
<polyline fill="none" stroke="#000000" points="1225.0433,-349.7123 1229.0433,-353.7123 "/>
<text text-anchor="middle" x="1073" y="-329.7123" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_page</text>
</g>
<!-- t12&#45;&gt;t15 -->
<g id="edge14" class="edge">
<title>t12&#45;&gt;t15</title>
<path fill="none" stroke="#0044a0" d="M754.3855,-411.2269C799.2649,-402.117 854.0781,-390.6067 903,-379.252 933.2642,-372.2277 966.3927,-363.8908 995.3012,-356.4015"/>
<polygon fill="#0044a0" stroke="#0044a0" points="996.5416,-359.6955 1005.3389,-353.7914 994.7799,-352.9208 996.5416,-359.6955"/>
</g>
<!-- t16 -->
<g id="node19" class="node">
<title>t16</title>
<polygon fill="#00cc66" stroke="#000000" points="651.9202,-270.1725 284.0798,-270.1725 280.0798,-266.1725 280.0798,-234.1725 647.9202,-234.1725 651.9202,-238.1725 651.9202,-270.1725"/>
<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 280.0798,-266.1725 "/>
<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 647.9202,-234.1725 "/>
<polyline fill="none" stroke="#000000" points="647.9202,-266.1725 651.9202,-270.1725 "/>
<text text-anchor="middle" x="466" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_debug_page</text>
</g>
<!-- t12&#45;&gt;t16 -->
<g id="edge27" class="edge">
<title>t12&#45;&gt;t16</title>
<path fill="none" stroke="#0044a0" d="M772.353,-411.2135C825.7002,-401.3017 880.0px85,-389.0648 888,-379.252 912.2169,-349.0628 914.0433,-320.8011 888,-292.1725 871.9898,-274.5731 764.4698,-264.5473 661.9779,-258.9344"/>
<polygon fill="#0044a0" stroke="#0044a0" points="661.9319,-255.4271 651.7593,-258.3883 661.5583,-262.4171 661.9319,-255.4271"/>
</g>
<!-- t25 -->
<g id="node27" class="node">
<title>t25</title>
<polygon fill="#efa03b" stroke="#000000" points="488.7205,-323.495 682,-296.1762 875.2795,-323.495 875.0991,-367.6976 488.9009,-367.6976 488.7205,-323.495"/>
<polygon fill="none" stroke="#000000" points="484.7033,-320.0193 682,-292.1327 879.2967,-320.0193 879.0858,-371.7013 484.9142,-371.7013 484.7033,-320.0193"/>
<text text-anchor="middle" x="682" y="-329.7123" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_sidecars</text>
</g>
<!-- t12&#45;&gt;t25 -->
<g id="edge52" class="edge">
<title>t12&#45;&gt;t25</title>
<path fill="none" stroke="#0044a0" d="M666.6634,-411.2166C668.3795,-402.768 670.5105,-392.2769 672.6326,-381.8293"/>
<polygon fill="#0044a0" stroke="#0044a0" points="676.09,-382.3905 674.6507,-371.8939 669.2301,-380.9971 676.09,-382.3905"/>
</g>
<!-- t18 -->
<g id="node18" class="node">
<title>t18</title>
<polygon fill="#00cc66" stroke="#000000" points="1147.2989,-270.1725 856.7011,-270.1725 852.7011,-266.1725 852.7011,-234.1725 1143.2989,-234.1725 1147.2989,-238.1725 1147.2989,-270.1725"/>
<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 852.7011,-266.1725 "/>
<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 1143.2989,-234.1725 "/>
<polyline fill="none" stroke="#000000" points="1143.2989,-266.1725 1147.2989,-270.1725 "/>
<text text-anchor="middle" x="1000" y="-246.1725" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.combine_layers</text>
</g>
<!-- t15&#45;&gt;t18 -->
<g id="edge24" class="edge">
<title>t15&#45;&gt;t18</title>
<path fill="none" stroke="#0044a0" d="M1057.1601,-317.5854C1047.0201,-305.9814 1033.7081,-290.7474 1022.4822,-277.9007"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1025.0247,-275.4911 1015.809,-270.264 1019.7536,-280.0972 1025.0247,-275.4911"/>
</g>
<!-- t17&#45;&gt;t18 -->
<g id="edge23" class="edge">
<title>t17&#45;&gt;t18</title>
<path fill="none" stroke="#0044a0" d="M1160.5767,-411.2248C1195.7121,-403.5317 1227.1378,-393.066 1238,-379.252 1261.9222,-348.8288 1262.9959,-321.72 1238,-292.1725 1231.6408,-284.6553 1197.8632,-277.2656 1157.6598,-270.9231"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1158.0117,-267.4362 1147.596,-269.3761 1156.9481,-274.3549 1158.0117,-267.4362"/>
</g>
<!-- t17&#45;&gt;t25 -->
<g id="edge51" class="edge">
<title>t17&#45;&gt;t25</title>
<path fill="none" stroke="#0044a0" d="M964.5165,-411.2166C925.304,-400.7368 874.3329,-387.1145 826.4989,-374.3305"/>
<polygon fill="#0044a0" stroke="#0044a0" points="827.2267,-370.9023 816.6621,-371.7016 825.4193,-377.6649 827.2267,-370.9023"/>
</g>
<!-- t13&#45;&gt;t14 -->
<g id="edge21" class="edge">
<title>t13&#45;&gt;t14</title>
<path fill="none" stroke="#0044a0" d="M672.8683,-493.1993C736.8314,-488.6166 808.4777,-483.5937 874,-479.252 1033.5051,-468.6828 1073.5671,-468.8587 1233,-457.252 1285.8429,-453.4051 1343.3743,-448.7264 1395.4741,-444.3037"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1395.81,-447.7879 1405.477,-443.4521 1395.2161,-440.8131 1395.81,-447.7879"/>
</g>
<!-- t13&#45;&gt;t16 -->
<g id="edge26" class="edge">
<title>t13&#45;&gt;t16</title>
<path fill="none" stroke="#0044a0" d="M479.4495,-479.3529C478.2545,-457.9761 476.5497,-427.7492 475,-401.252 472.546,-359.2945 469.5914,-310.7019 467.754,-280.6981"/>
<polygon fill="#0044a0" stroke="#0044a0" points="471.2362,-280.2988 467.1306,-270.5318 464.2493,-280.7273 471.2362,-280.2988"/>
</g>
<!-- t13&#45;&gt;t19 -->
<g id="edge29" class="edge">
<title>t13&#45;&gt;t19</title>
<path fill="none" stroke="#0044a0" d="M415.2907,-483.9551C385.6007,-473.4286 350.7427,-461.0699 321.6782,-450.7652"/>
<polygon fill="#0044a0" stroke="#0044a0" points="322.6542,-447.3978 312.0595,-447.3549 320.315,-453.9954 322.6542,-447.3978"/>
</g>
<!-- t14&#45;&gt;t18 -->
<g id="edge25" class="edge">
<title>t14&#45;&gt;t18</title>
<path fill="none" stroke="#0044a0" d="M1519.9776,-405.138C1460.3688,-374.2505 1351.1978,-321.2463 1252,-292.1725 1221.8888,-283.3473 1189.1477,-276.3434 1157.5373,-270.8287"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1157.7641,-267.3171 1147.3187,-269.0901 1156.59,-274.2179 1157.7641,-267.3171"/>
<path fill="none" stroke="#0044a0" d="M1704,-1121.0193C1704,-1097.893 1704,-1057.7125 1704,-1030.5603"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1707.5001,-1030.5066 1704,-1020.5066 1700.5001,-1030.5067 1707.5001,-1030.5066"/>
</g>
<!-- t22 -->
<g id="node23" class="node">
<title>t22</title>
<polygon fill="#efa03b" stroke="#000000" points="489.918,-156.4263 735,-129.1169 980.082,-156.4263 979.8532,-200.6137 490.1468,-200.6137 489.918,-156.4263"/>
<polygon fill="none" stroke="#000000" points="485.9027,-152.8467 735,-125.09 984.0973,-152.8467 983.8293,-204.6159 486.1707,-204.6159 485.9027,-152.8467"/>
<text text-anchor="middle" x="735" y="-162.6394" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages_ghostscript</text>
<polygon fill="#efa03b" stroke="#000000" points="1542.1753,-156.4327 1723,-129.1104 1903.8247,-156.4327 1903.6559,-200.6411 1542.3441,-200.6411 1542.1753,-156.4327"/>
<polygon fill="none" stroke="#000000" points="1538.1584,-152.99 1723,-125.0608 1907.8416,-152.99 1907.6444,-204.6453 1538.3556,-204.6453 1538.1584,-152.99"/>
<text text-anchor="middle" x="1723" y="-162.6516" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_pages</text>
</g>
<!-- t1&#45;&gt;t22 -->
<g id="edge37" class="edge">
<title>t1&#45;&gt;t22</title>
<path fill="none" stroke="#0044a0" d="M1779.013,-1121.1962C1836.8401,-1102.3174 1907,-1066.2281 1907,-1002.2438 1907,-1002.2438 1907,-1002.2438 1907,-335.7367 1907,-285.0344 1892.7171,-268.0813 1855,-234.197 1844.8914,-225.1156 1833.2959,-216.9812 1821.2655,-209.7918"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1822.8581,-206.6703 1812.4436,-204.7237 1819.3711,-212.74 1822.8581,-206.6703"/>
</g>
<!-- t3 -->
<g id="node4" class="node">
<title>t3</title>
<polygon fill="#efa03b" stroke="#000000" points="1502.3851,-1020.2438 1255.6149,-1020.2438 1251.6149,-1016.2438 1251.6149,-984.2438 1498.3851,-984.2438 1502.3851,-988.2438 1502.3851,-1020.2438"/>
<polyline fill="none" stroke="#000000" points="1498.3851,-1016.2438 1251.6149,-1016.2438 "/>
<polyline fill="none" stroke="#000000" points="1498.3851,-1016.2438 1498.3851,-984.2438 "/>
<polyline fill="none" stroke="#000000" points="1498.3851,-1016.2438 1502.3851,-1020.2438 "/>
<text text-anchor="middle" x="1377" y="-996.2438" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.split_page</text>
</g>
<!-- t2&#45;&gt;t3 -->
<g id="edge3" class="edge">
<title>t2&#45;&gt;t3</title>
<path fill="none" stroke="#0044a0" d="M1391.9557,-1046.7738C1390.1421,-1041.3739 1388.2021,-1035.5977 1386.3405,-1030.0546"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1389.62,-1028.826 1383.1183,-1020.4607 1382.9843,-1031.0547 1389.62,-1028.826"/>
</g>
<!-- t4 -->
<g id="node5" class="node">
<title>t4</title>
<polygon fill="#efa03b" stroke="#000000" points="1607.5194,-941.3183 1377,-958.219 1146.4806,-941.3183 1146.6958,-913.9725 1607.3042,-913.9725 1607.5194,-941.3183"/>
<polygon fill="none" stroke="#000000" points="1611.5491,-945.0313 1377,-962.2274 1142.4509,-945.0313 1142.7267,-909.9748 1611.2733,-909.9748 1611.5491,-945.0313"/>
<text text-anchor="middle" x="1377" y="-927.7601" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_or_skip</text>
</g>
<!-- t3&#45;&gt;t4 -->
<g id="edge4" class="edge">
<title>t3&#45;&gt;t4</title>
<path fill="none" stroke="#0044a0" d="M1377,-983.8914C1377,-980.3974 1377,-976.6197 1377,-972.7598"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1380.5001,-972.4886 1377,-962.4886 1373.5001,-972.4887 1380.5001,-972.4886"/>
</g>
<!-- t5 -->
<g id="node6" class="node">
<title>t5</title>
<polygon fill="#efa03b" stroke="#000000" points="1405.3166,-883.2765 1098.6834,-883.2765 1094.6834,-879.2765 1094.6834,-847.2765 1401.3166,-847.2765 1405.3166,-851.2765 1405.3166,-883.2765"/>
<polyline fill="none" stroke="#000000" points="1401.3166,-879.2765 1094.6834,-879.2765 "/>
<polyline fill="none" stroke="#000000" points="1401.3166,-879.2765 1401.3166,-847.2765 "/>
<polyline fill="none" stroke="#000000" points="1401.3166,-879.2765 1405.3166,-883.2765 "/>
<text text-anchor="middle" x="1250" y="-859.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_preview</text>
</g>
<!-- t4&#45;&gt;t5 -->
<g id="edge5" class="edge">
<title>t4&#45;&gt;t5</title>
<path fill="none" stroke="#0044a0" d="M1332.5816,-909.8079C1319.5732,-902.7932 1305.3872,-895.1435 1292.5011,-888.1948"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1293.8464,-884.9439 1283.3833,-883.2781 1290.5239,-891.1052 1293.8464,-884.9439"/>
</g>
<!-- t6 -->
<g id="node7" class="node">
<title>t6</title>
<polygon fill="#efa03b" stroke="#000000" points="1475.0824,-825.2765 1216.9176,-825.2765 1212.9176,-821.2765 1212.9176,-789.2765 1471.0824,-789.2765 1475.0824,-793.2765 1475.0824,-825.2765"/>
<polyline fill="none" stroke="#000000" points="1471.0824,-821.2765 1212.9176,-821.2765 "/>
<polyline fill="none" stroke="#000000" points="1471.0824,-821.2765 1471.0824,-789.2765 "/>
<polyline fill="none" stroke="#000000" points="1471.0824,-821.2765 1475.0824,-825.2765 "/>
<text text-anchor="middle" x="1344" y="-801.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.orient_page</text>
</g>
<!-- t4&#45;&gt;t6 -->
<g id="edge7" class="edge">
<title>t4&#45;&gt;t6</title>
<path fill="none" stroke="#0044a0" d="M1399.1144,-909.8909C1412.803,-892.0931 1425.7216,-867.4749 1414,-847.2765 1410.4598,-841.1761 1405.6012,-835.9275 1400.0802,-831.4317"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1401.7661,-828.3348 1391.594,-825.372 1397.6983,-834.0316 1401.7661,-828.3348"/>
</g>
<!-- t5&#45;&gt;t6 -->
<g id="edge6" class="edge">
<title>t5&#45;&gt;t6</title>
<path fill="none" stroke="#0044a0" d="M1279.2269,-847.2428C1287.7153,-842.0053 1297.1107,-836.2082 1306.0518,-830.6913"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1308.1187,-833.5287 1314.7912,-825.2989 1304.4429,-827.5714 1308.1187,-833.5287"/>
</g>
<!-- t7 -->
<g id="node8" class="node">
<title>t7</title>
<polygon fill="#efa03b" stroke="#000000" points="1533.7325,-767.2765 1158.2675,-767.2765 1154.2675,-763.2765 1154.2675,-731.2765 1529.7325,-731.2765 1533.7325,-735.2765 1533.7325,-767.2765"/>
<polyline fill="none" stroke="#000000" points="1529.7325,-763.2765 1154.2675,-763.2765 "/>
<polyline fill="none" stroke="#000000" points="1529.7325,-763.2765 1529.7325,-731.2765 "/>
<polyline fill="none" stroke="#000000" points="1529.7325,-763.2765 1533.7325,-767.2765 "/>
<text text-anchor="middle" x="1344" y="-743.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.rasterize_with_ghostscript</text>
</g>
<!-- t6&#45;&gt;t7 -->
<g id="edge8" class="edge">
<title>t6&#45;&gt;t7</title>
<path fill="none" stroke="#0044a0" d="M1344,-789.2428C1344,-785.5683 1344,-781.6182 1344,-777.6828"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1347.5001,-777.4987 1344,-767.4987 1340.5001,-777.4987 1347.5001,-777.4987"/>
</g>
<!-- t17 -->
<g id="node15" class="node">
<title>t17</title>
<polygon fill="#ff69b4" stroke="#000000" points="1064.2696,-447.2765 683.7304,-447.2765 679.7304,-443.2765 679.7304,-411.2765 1060.2696,-411.2765 1064.2696,-415.2765 1064.2696,-447.2765"/>
<polyline fill="none" stroke="#000000" points="1060.2696,-443.2765 679.7304,-443.2765 "/>
<polyline fill="none" stroke="#000000" points="1060.2696,-443.2765 1060.2696,-411.2765 "/>
<polyline fill="none" stroke="#000000" points="1060.2696,-443.2765 1064.2696,-447.2765 "/>
<text text-anchor="middle" x="872" y="-423.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_textonly_pdf</text>
</g>
<!-- t6&#45;&gt;t17 -->
<g id="edge16" class="edge">
<title>t6&#45;&gt;t17</title>
<path fill="none" stroke="#0044a0" d="M1212.9072,-798.053C1014.8011,-782.3275 662,-746.5615 662,-691.2765 662,-691.2765 662,-691.2765 662,-575.2765 662,-532.3789 643.882,-511.6736 672,-479.2765 682.6302,-467.0285 710.7979,-457.1862 742.7543,-449.5814"/>
<polygon fill="#0044a0" stroke="#0044a0" points="743.8081,-452.9309 752.7782,-447.2929 742.25,-446.1065 743.8081,-452.9309"/>
</g>
<!-- t14 -->
<g id="node17" class="node">
<title>t14</title>
<polygon fill="#00cc66" stroke="#000000" points="339,-457.2765 16.4141,-429.2765 339,-401.2765 661.5859,-429.2765 339,-457.2765"/>
<text text-anchor="middle" x="339" y="-423.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_image_layer</text>
</g>
<!-- t6&#45;&gt;t14 -->
<g id="edge22" class="edge">
<title>t6&#45;&gt;t14</title>
<path fill="none" stroke="#0044a0" d="M1212.7923,-802.9902C970.4193,-793.3094 477,-764.5974 477,-691.2765 477,-691.2765 477,-691.2765 477,-575.2765 477,-524.368 432.1386,-484.0495 393.5722,-458.7944"/>
<polygon fill="#0044a0" stroke="#0044a0" points="395.2709,-455.7271 384.952,-453.3247 391.5205,-461.6377 395.2709,-455.7271"/>
</g>
<!-- t21 -->
<g id="node20" class="node">
<title>t21</title>
<polygon fill="#efa03b" stroke="#000000" points="1867.7716,-709.2765 1622.2284,-709.2765 1618.2284,-705.2765 1618.2284,-673.2765 1863.7716,-673.2765 1867.7716,-677.2765 1867.7716,-709.2765"/>
<polyline fill="none" stroke="#000000" points="1863.7716,-705.2765 1618.2284,-705.2765 "/>
<polyline fill="none" stroke="#000000" points="1863.7716,-705.2765 1863.7716,-673.2765 "/>
<polyline fill="none" stroke="#000000" points="1863.7716,-705.2765 1867.7716,-709.2765 "/>
<text text-anchor="middle" x="1743" y="-685.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.skip_page</text>
</g>
<!-- t6&#45;&gt;t21 -->
<g id="edge28" class="edge">
<title>t6&#45;&gt;t21</title>
<path fill="none" stroke="#0044a0" d="M1475.3964,-796.0838C1517.2386,-790.1439 1563.2301,-781.082 1604,-767.2765 1641.4476,-754.5959 1680.6677,-732.1928 1707.8621,-714.9842"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1710.0816,-717.7185 1716.6036,-709.369 1706.2983,-711.8289 1710.0816,-717.7185"/>
</g>
<!-- t19 -->
<g id="node21" class="node">
<title>t19</title>
<polygon fill="#66ccff" stroke="#000000" points="1847.6739,-447.2765 1442.3261,-447.2765 1438.3261,-443.2765 1438.3261,-411.2765 1843.6739,-411.2765 1847.6739,-415.2765 1847.6739,-447.2765"/>
<polyline fill="none" stroke="#000000" points="1843.6739,-443.2765 1438.3261,-443.2765 "/>
<polyline fill="none" stroke="#000000" points="1843.6739,-443.2765 1843.6739,-411.2765 "/>
<polyline fill="none" stroke="#000000" points="1843.6739,-443.2765 1847.6739,-447.2765 "/>
<text text-anchor="middle" x="1643" y="-423.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_and_render_pdf</text>
</g>
<!-- t6&#45;&gt;t19 -->
<g id="edge30" class="edge">
<title>t6&#45;&gt;t19</title>
<path fill="none" stroke="#0044a0" d="M1471.3274,-789.2627C1502.1321,-783.2141 1529.9438,-775.8026 1543,-767.2765 1585.741,-739.3653 1573.921,-710.3614 1609,-673.2765 1677.0666,-601.3175 1740.7796,-624.4008 1784,-535.2765 1794.8601,-512.8819 1798.3601,-499.6049 1784,-479.2765 1776.4259,-468.5545 1752.8487,-458.504 1727.0874,-450.2665"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1728.1106,-446.9194 1717.5232,-447.3158 1726.0469,-453.6083 1728.1106,-446.9194"/>
</g>
<!-- t8 -->
<g id="node9" class="node">
<title>t8</title>
<polygon fill="#efa03b" stroke="#000000" points="1418.0801,-709.2765 993.9199,-709.2765 989.9199,-705.2765 989.9199,-673.2765 1414.0801,-673.2765 1418.0801,-677.2765 1418.0801,-709.2765"/>
<polyline fill="none" stroke="#000000" points="1414.0801,-705.2765 989.9199,-705.2765 "/>
<polyline fill="none" stroke="#000000" points="1414.0801,-705.2765 1414.0801,-673.2765 "/>
<polyline fill="none" stroke="#000000" points="1414.0801,-705.2765 1418.0801,-709.2765 "/>
<text text-anchor="middle" x="1204" y="-685.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_remove_background</text>
</g>
<!-- t7&#45;&gt;t8 -->
<g id="edge9" class="edge">
<title>t7&#45;&gt;t8</title>
<path fill="none" stroke="#0044a0" d="M1300.4705,-731.2428C1286.6431,-725.5143 1271.1996,-719.1163 1256.7928,-713.1478"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1258.0807,-709.8929 1247.5025,-709.2989 1255.4015,-716.3599 1258.0807,-709.8929"/>
</g>
<!-- t13 -->
<g id="node16" class="node">
<title>t13</title>
<polygon fill="#efa03b" stroke="#000000" points="1391,-535.2765 1007.1874,-507.2765 1391,-479.2765 1774.8126,-507.2765 1391,-535.2765"/>
<text text-anchor="middle" x="1391" y="-501.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_visible_page_image</text>
</g>
<!-- t7&#45;&gt;t13 -->
<g id="edge20" class="edge">
<title>t7&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M1403.1818,-731.1462C1412.6461,-725.627 1421.2143,-718.4872 1427,-709.2765 1459.1275,-658.1296 1432.9036,-586.12 1411.5309,-543.3218"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1414.6214,-541.6783 1406.9295,-534.3924 1408.399,-544.8848 1414.6214,-541.6783"/>
</g>
<!-- t9 -->
<g id="node10" class="node">
<title>t9</title>
<polygon fill="#efa03b" stroke="#000000" points="1365.6759,-651.2765 1046.3241,-651.2765 1042.3241,-647.2765 1042.3241,-615.2765 1361.6759,-615.2765 1365.6759,-619.2765 1365.6759,-651.2765"/>
<polyline fill="none" stroke="#000000" points="1361.6759,-647.2765 1042.3241,-647.2765 "/>
<polyline fill="none" stroke="#000000" points="1361.6759,-647.2765 1361.6759,-615.2765 "/>
<polyline fill="none" stroke="#000000" points="1361.6759,-647.2765 1365.6759,-651.2765 "/>
<text text-anchor="middle" x="1204" y="-627.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_deskew</text>
</g>
<!-- t8&#45;&gt;t9 -->
<g id="edge10" class="edge">
<title>t8&#45;&gt;t9</title>
<path fill="none" stroke="#0044a0" d="M1204,-673.2428C1204,-669.5683 1204,-665.6182 1204,-661.6828"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1207.5001,-661.4987 1204,-651.4987 1200.5001,-661.4987 1207.5001,-661.4987"/>
</g>
<!-- t8&#45;&gt;t13 -->
<g id="edge19" class="edge">
<title>t8&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M1324.7334,-673.2711C1347.6618,-667.5415 1367.0063,-660.2841 1375,-651.2765 1400.3856,-622.6711 1401.2608,-577.249 1397.7857,-545.0765"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1401.2555,-544.6171 1396.5301,-535.1345 1394.3107,-545.4942 1401.2555,-544.6171"/>
</g>
<!-- t10 -->
<g id="node11" class="node">
<title>t10</title>
<polygon fill="#efa03b" stroke="#000000" points="1351.8828,-593.2765 1050.1172,-593.2765 1046.1172,-589.2765 1046.1172,-557.2765 1347.8828,-557.2765 1351.8828,-561.2765 1351.8828,-593.2765"/>
<polyline fill="none" stroke="#000000" points="1347.8828,-589.2765 1046.1172,-589.2765 "/>
<polyline fill="none" stroke="#000000" points="1347.8828,-589.2765 1347.8828,-557.2765 "/>
<polyline fill="none" stroke="#000000" points="1347.8828,-589.2765 1351.8828,-593.2765 "/>
<text text-anchor="middle" x="1199" y="-569.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.preprocess_clean</text>
</g>
<!-- t9&#45;&gt;t10 -->
<g id="edge11" class="edge">
<title>t9&#45;&gt;t10</title>
<path fill="none" stroke="#0044a0" d="M1202.4454,-615.2428C1202.1286,-611.5683 1201.7881,-607.6182 1201.4488,-603.6828"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1204.9169,-603.1611 1200.5709,-593.4987 1197.9428,-603.7624 1204.9169,-603.1611"/>
</g>
<!-- t9&#45;&gt;t13 -->
<g id="edge18" class="edge">
<title>t9&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M1305.4912,-615.2209C1329.9901,-609.1504 1351.7763,-601.7378 1361,-593.2765 1374.4133,-580.9719 1381.8295,-562.3451 1385.93,-545.5694"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1389.4281,-545.945 1388.0785,-535.4364 1382.5803,-544.4929 1389.4281,-545.945"/>
</g>
<!-- t11 -->
<g id="node12" class="node">
<title>t11</title>
<polygon fill="#efa03b" stroke="#000000" points="989.1094,-525.2765 684.8906,-525.2765 680.8906,-521.2765 680.8906,-489.2765 985.1094,-489.2765 989.1094,-493.2765 989.1094,-525.2765"/>
<polyline fill="none" stroke="#000000" points="985.1094,-521.2765 680.8906,-521.2765 "/>
<polyline fill="none" stroke="#000000" points="985.1094,-521.2765 985.1094,-489.2765 "/>
<polyline fill="none" stroke="#000000" points="985.1094,-521.2765 989.1094,-525.2765 "/>
<text text-anchor="middle" x="835" y="-501.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.select_ocr_image</text>
</g>
<!-- t10&#45;&gt;t11 -->
<g id="edge12" class="edge">
<title>t10&#45;&gt;t11</title>
<path fill="none" stroke="#0044a0" d="M1102.4103,-557.2322C1052.8491,-547.9736 992.3597,-536.6733 941.4558,-527.1638"/>
<polygon fill="#0044a0" stroke="#0044a0" points="941.8816,-523.6829 931.4089,-525.2869 940.5961,-530.5638 941.8816,-523.6829"/>
</g>
<!-- t10&#45;&gt;t13 -->
<g id="edge17" class="edge">
<title>t10&#45;&gt;t13</title>
<path fill="none" stroke="#0044a0" d="M1249.9484,-557.2322C1270.0775,-550.1032 1293.6242,-541.7637 1315.6488,-533.9633"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1317.0528,-537.1792 1325.3106,-530.5415 1314.7158,-530.5808 1317.0528,-537.1792"/>
</g>
<!-- t12 -->
<g id="node13" class="node">
<title>t12</title>
<polygon fill="#00cc66" stroke="#000000" points="1400.0996,-447.2765 1085.9004,-447.2765 1081.9004,-443.2765 1081.9004,-411.2765 1396.0996,-411.2765 1400.0996,-415.2765 1400.0996,-447.2765"/>
<polyline fill="none" stroke="#000000" points="1396.0996,-443.2765 1081.9004,-443.2765 "/>
<polyline fill="none" stroke="#000000" points="1396.0996,-443.2765 1396.0996,-411.2765 "/>
<polyline fill="none" stroke="#000000" points="1396.0996,-443.2765 1400.0996,-447.2765 "/>
<text text-anchor="middle" x="1241" y="-423.2765" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.ocr_tesseract_hocr</text>
</g>
<!-- t11&#45;&gt;t12 -->
<g id="edge13" class="edge">
<title>t11&#45;&gt;t12</title>
<path fill="none" stroke="#0044a0" d="M928.9096,-489.2347C991.1602,-477.2753 1072.8804,-461.5753 1137.0821,-449.241"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1138.07,-452.6153 1147.23,-447.2914 1136.7492,-445.741 1138.07,-452.6153"/>
</g>
<!-- t11&#45;&gt;t17 -->
<g id="edge15" class="edge">
<title>t11&#45;&gt;t17</title>
<path fill="none" stroke="#0044a0" d="M843.5815,-489.1856C848.1717,-479.5091 853.9184,-467.3944 859.0408,-456.5958"/>
<polygon fill="#0044a0" stroke="#0044a0" points="862.2132,-458.0745 863.3368,-447.5394 855.8887,-455.0743 862.2132,-458.0745"/>
</g>
<!-- t15 -->
<g id="node14" class="node">
<title>t15</title>
<polygon fill="#00cc66" stroke="#000000" points="1398.0433,-353.7367 1089.9567,-353.7367 1085.9567,-349.7367 1085.9567,-317.7367 1394.0433,-317.7367 1398.0433,-321.7367 1398.0433,-353.7367"/>
<polyline fill="none" stroke="#000000" points="1394.0433,-349.7367 1085.9567,-349.7367 "/>
<polyline fill="none" stroke="#000000" points="1394.0433,-349.7367 1394.0433,-317.7367 "/>
<polyline fill="none" stroke="#000000" points="1394.0433,-349.7367 1398.0433,-353.7367 "/>
<text text-anchor="middle" x="1242" y="-329.7367" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_page</text>
</g>
<!-- t12&#45;&gt;t15 -->
<g id="edge14" class="edge">
<title>t12&#45;&gt;t15</title>
<path fill="none" stroke="#0044a0" d="M1241.1928,-411.2411C1241.3358,-397.8693 1241.5334,-379.3811 1241.6973,-364.0466"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1245.1972,-364.0728 1241.8044,-354.0359 1238.1976,-363.9979 1245.1972,-364.0728"/>
</g>
<!-- t16 -->
<g id="node19" class="node">
<title>t16</title>
<polygon fill="#00cc66" stroke="#000000" points="1787.9202,-353.7367 1420.0798,-353.7367 1416.0798,-349.7367 1416.0798,-317.7367 1783.9202,-317.7367 1787.9202,-321.7367 1787.9202,-353.7367"/>
<polyline fill="none" stroke="#000000" points="1783.9202,-349.7367 1416.0798,-349.7367 "/>
<polyline fill="none" stroke="#000000" points="1783.9202,-349.7367 1783.9202,-317.7367 "/>
<polyline fill="none" stroke="#000000" points="1783.9202,-349.7367 1787.9202,-353.7367 "/>
<text text-anchor="middle" x="1602" y="-329.7367" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.render_hocr_debug_page</text>
</g>
<!-- t12&#45;&gt;t16 -->
<g id="edge27" class="edge">
<title>t12&#45;&gt;t16</title>
<path fill="none" stroke="#0044a0" d="M1310.6044,-411.2411C1371.1859,-395.5436 1458.9787,-372.7954 1522.3491,-356.3753"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1523.4996,-359.6928 1532.302,-353.7964 1521.7438,-352.9166 1523.4996,-359.6928"/>
</g>
<!-- t23 -->
<g id="node25" class="node">
<title>t23</title>
<polygon fill="#efa03b" stroke="#000000" points="677.7205,-323.5194 871,-296.2006 1064.2795,-323.5194 1064.0991,-367.7221 677.9009,-367.7221 677.7205,-323.5194"/>
<polygon fill="none" stroke="#000000" points="673.7033,-320.0438 871,-292.1572 1068.2967,-320.0438 1068.0858,-371.7257 673.9142,-371.7257 673.7033,-320.0438"/>
<text text-anchor="middle" x="871" y="-329.7367" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.merge_sidecars</text>
</g>
<!-- t12&#45;&gt;t23 -->
<g id="edge41" class="edge">
<title>t12&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M1169.6603,-411.2411C1128.034,-400.7175 1073.8731,-387.0251 1023.1228,-374.1949"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1023.9099,-370.7839 1013.3571,-371.726 1022.1942,-377.5703 1023.9099,-370.7839"/>
</g>
<!-- t18 -->
<g id="node18" class="node">
<title>t18</title>
<polygon fill="#00cc66" stroke="#000000" points="1096.2989,-270.197 805.7011,-270.197 801.7011,-266.197 801.7011,-234.197 1092.2989,-234.197 1096.2989,-238.197 1096.2989,-270.197"/>
<polyline fill="none" stroke="#000000" points="1092.2989,-266.197 801.7011,-266.197 "/>
<polyline fill="none" stroke="#000000" points="1092.2989,-266.197 1092.2989,-234.197 "/>
<polyline fill="none" stroke="#000000" points="1092.2989,-266.197 1096.2989,-270.197 "/>
<text text-anchor="middle" x="949" y="-246.197" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.combine_layers</text>
</g>
<!-- t15&#45;&gt;t18 -->
<g id="edge24" class="edge">
<title>t15&#45;&gt;t18</title>
<path fill="none" stroke="#0044a0" d="M1178.7841,-317.7127C1132.7483,-304.587 1070.2826,-286.7769 1022.1,-273.0391"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1022.9314,-269.6368 1012.355,-270.2607 1021.012,-276.3685 1022.9314,-269.6368"/>
</g>
<!-- t17&#45;&gt;t18 -->
<g id="edge23" class="edge">
<title>t17&#45;&gt;t18</title>
<path fill="none" stroke="#0044a0" d="M742.5275,-411.2563C707.334,-403.5731 675.8911,-393.1114 665,-379.2765 641.0609,-348.8665 639.6136,-321.4095 665,-292.197 674.2623,-281.5388 731.0775,-272.4723 791.4852,-265.6796"/>
<polygon fill="#0044a0" stroke="#0044a0" points="792.0426,-269.1394 801.5997,-264.5664 791.2768,-262.1814 792.0426,-269.1394"/>
</g>
<!-- t17&#45;&gt;t23 -->
<g id="edge40" class="edge">
<title>t17&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M871.8072,-411.2411C871.7178,-402.8795 871.607,-392.5175 871.4965,-382.1769"/>
<polygon fill="#0044a0" stroke="#0044a0" points="874.9936,-381.8803 871.3868,-371.9184 867.994,-381.9552 874.9936,-381.8803"/>
</g>
<!-- t13&#45;&gt;t14 -->
<g id="edge21" class="edge">
<title>t13&#45;&gt;t14</title>
<path fill="none" stroke="#0044a0" d="M1197.8658,-493.3637C1134.2194,-488.8196 1063.0826,-483.7842 998,-479.2765 852.6863,-469.2117 816.2452,-468.2864 671,-457.2765 618.6842,-453.3108 561.7627,-448.6403 510.0667,-444.2623"/>
<polygon fill="#0044a0" stroke="#0044a0" points="510.0603,-440.7492 499.8001,-443.3908 509.4682,-447.7241 510.0603,-440.7492"/>
</g>
<!-- t13&#45;&gt;t16 -->
<g id="edge26" class="edge">
<title>t13&#45;&gt;t16</title>
<path fill="none" stroke="#0044a0" d="M1400.9154,-479.7728C1403.5471,-472.4624 1406.3842,-464.5715 1409,-457.2765 1417.9204,-432.3989 1410.4916,-420.142 1429,-401.2765 1448.8322,-381.0616 1475.5255,-366.9968 1502.0009,-357.2438"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1503.4471,-360.446 1511.7321,-353.8422 1501.1373,-353.8381 1503.4471,-360.446"/>
</g>
<!-- t13&#45;&gt;t19 -->
<g id="edge29" class="edge">
<title>t13&#45;&gt;t19</title>
<path fill="none" stroke="#0044a0" d="M1464.2273,-484.6109C1499.1449,-473.803 1540.6621,-460.9525 1574.8943,-450.3568"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1576.2874,-453.5895 1584.8054,-447.2891 1574.2176,-446.9025 1576.2874,-453.5895"/>
</g>
<!-- t14&#45;&gt;t18 -->
<g id="edge25" class="edge">
<title>t14&#45;&gt;t18</title>
<path fill="none" stroke="#0044a0" d="M382.1114,-404.8262C439.5148,-373.4449 545.0952,-319.7698 642,-292.197 689.7605,-278.6074 743.0314,-269.6169 791.6153,-263.6758"/>
<polygon fill="#0044a0" stroke="#0044a0" points="792.0696,-267.1466 801.5866,-262.4909 791.2434,-260.1955 792.0696,-267.1466"/>
</g>
<!-- t18&#45;&gt;t22 -->
<g id="edge35" class="edge">
<title>t18&#45;&gt;t22</title>
<path fill="none" stroke="#0044a0" d="M942.8252,-234.1499C918.2191,-226.3936 888.4232,-217.0013 859.1708,-207.7804"/>
<polygon fill="#0044a0" stroke="#0044a0" points="859.8692,-204.3309 849.2796,-204.6625 857.7647,-211.007 859.8692,-204.3309"/>
</g>
<!-- t18&#45;&gt;t24 -->
<g id="edge39" class="edge">
<title>t18&#45;&gt;t24</title>
<path fill="none" stroke="#0044a0" d="M852.4716,-245.1051C747.4645,-239.0474 603.1919,-228.6387 477,-212.1725 463.5912,-210.4229 449.788,-208.4132 435.908,-206.238"/>
<polygon fill="#0044a0" stroke="#0044a0" points="436.2183,-202.7434 425.7921,-204.6263 435.1168,-209.6562 436.2183,-202.7434"/>
</g>
<!-- t18&#45;&gt;t23 -->
<g id="edge44" class="edge">
<title>t18&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M1047.0344,-234.1499C1066.9243,-226.5285 1090.9364,-217.3275 1114.596,-208.2617"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1115.9033,-211.509 1123.9889,-204.6625 1113.3986,-204.9724 1115.9033,-211.509"/>
<path fill="none" stroke="#0044a0" d="M1096.4053,-236.2861C1218.8926,-223.0648 1394.1726,-204.1451 1528.216,-189.6765"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1528.6338,-193.1519 1538.2004,-188.5988 1527.8825,-186.1923 1528.6338,-193.1519"/>
</g>
<!-- t16&#45;&gt;t22 -->
<g id="edge36" class="edge">
<title>t16&#45;&gt;t22</title>
<path fill="none" stroke="#0044a0" d="M524.0378,-234.1499C549.1239,-226.3599 579.5241,-216.9196 609.3422,-207.6602"/>
<polygon fill="#0044a0" stroke="#0044a0" points="610.4833,-210.9707 618.9955,-204.6625 608.4073,-204.2857 610.4833,-210.9707"/>
</g>
<!-- t16&#45;&gt;t24 -->
<g id="edge40" class="edge">
<title>t16&#45;&gt;t24</title>
<path fill="none" stroke="#0044a0" d="M417.6711,-234.1499C397.2337,-226.5285 372.5607,-217.3275 348.2499,-208.2617"/>
<polygon fill="#0044a0" stroke="#0044a0" points="349.1912,-204.8773 338.5986,-204.6625 346.7453,-211.4361 349.1912,-204.8773"/>
</g>
<!-- t16&#45;&gt;t23 -->
<g id="edge45" class="edge">
<title>t16&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M651.7596,-235.3105C656.5547,-234.9215 661.308,-234.5415 666,-234.1725 811.2137,-222.7545 848.5512,-230.9322 993,-212.1725 1006.176,-210.4614 1019.7432,-208.4558 1033.3753,-206.2619"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1034.0068,-209.7051 1043.3083,-204.6325 1032.8736,-202.7974 1034.0068,-209.7051"/>
<path fill="none" stroke="#0044a0" d="M1615.1795,-317.5376C1633.1921,-292.6646 1666.0831,-247.2464 1690.8186,-213.0898"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1693.887,-214.8201 1696.9176,-204.668 1688.2175,-210.7143 1693.887,-214.8201"/>
</g>
<!-- t21&#45;&gt;t22 -->
<g id="edge32" class="edge">
<title>t21&#45;&gt;t22</title>
<path fill="none" stroke="#0044a0" d="M1165.0941,-235.3973C1085.9035,-224.7541 997.501,-212.8552 993,-212.1725 980.474,-210.2728 967.615,-208.2703 954.6442,-206.2099"/>
<polygon fill="#0044a0" stroke="#0044a0" points="955.1339,-202.7439 944.7072,-204.6238 954.0305,-209.6564 955.1339,-202.7439"/>
</g>
<!-- t21&#45;&gt;t24 -->
<g id="edge37" class="edge">
<title>t21&#45;&gt;t24</title>
<path fill="none" stroke="#0044a0" d="M1165.2215,-234.9906C1162.1187,-234.699 1159.041,-234.4254 1156,-234.1725 855.1023,-209.1529 777.1841,-244.6523 477,-212.1725 462.026,-210.5524 446.5737,-208.5164 431.0798,-206.2151"/>
<polygon fill="#0044a0" stroke="#0044a0" points="431.3094,-202.7099 420.8968,-204.6663 430.2567,-209.6303 431.3094,-202.7099"/>
</g>
<!-- t21&#45;&gt;t23 -->
<g id="edge42" class="edge">
<title>t21&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M1274.3771,-234.0471C1268.8557,-227.6412 1262.3807,-220.0px91 1255.8248,-212.5231"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1258.4543,-210.2128 1249.2743,-204.9233 1253.152,-214.783 1258.4543,-210.2128"/>
<path fill="none" stroke="#0044a0" d="M1784.6352,-673.2357C1820.6585,-654.4909 1867,-621.4113 1867,-575.2765 1867,-575.2765 1867,-575.2765 1867,-335.7367 1867,-284.2548 1827.2878,-240.7654 1789.5729,-211.0717"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1791.6387,-208.2453 1781.5703,-204.9469 1787.3842,-213.8041 1791.6387,-208.2453"/>
</g>
<!-- t19&#45;&gt;t22 -->
<g id="edge34" class="edge">
<title>t19&#45;&gt;t22</title>
<path fill="none" stroke="#0044a0" d="M260.6159,-410.9505C259.8316,-364.871 259.0249,-246.8273 271,-234.1725 302.6435,-200.7331 431.3919,-218.4498 477,-212.1725 490.4888,-210.316 504.3558,-208.2974 518.3278,-206.1801"/>
<polygon fill="#0044a0" stroke="#0044a0" points="519.1602,-209.5936 528.5161,-204.6219 518.1019,-202.6741 519.1602,-209.5936"/>
</g>
<!-- t19&#45;&gt;t24 -->
<g id="edge38" class="edge">
<title>t19&#45;&gt;t24</title>
<path fill="none" stroke="#0044a0" d="M259.0498,-411.2136C258.0595,-401.7112 256.8826,-389.8585 256,-379.252 251.2976,-322.7418 247.1648,-257.5963 244.6312,-214.9019"/>
<polygon fill="#0044a0" stroke="#0044a0" points="248.1192,-214.5935 244.0377,-204.8165 241.1313,-215.0048 248.1192,-214.5935"/>
<path fill="none" stroke="#0044a0" d="M1727.2665,-411.2415C1758.1998,-402.7736 1787.9112,-391.7833 1797,-379.2765 1833.7864,-328.6561 1794.7398,-258.1105 1760.8796,-212.906"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1763.5015,-210.5735 1754.637,-204.7708 1757.9481,-214.835 1763.5015,-210.5735"/>
</g>
<!-- t19&#45;&gt;t23 -->
<g id="edge43" class="edge">
<g id="edge39" class="edge">
<title>t19&#45;&gt;t23</title>
<path fill="none" stroke="#0044a0" d="M282.1528,-411.0561C318.8331,-380.6455 397.8323,-320.0855 476,-292.1725 553.979,-264.3269 579.2078,-283.062 661,-270.1725 742.8817,-257.269 762.3086,-248.2304 844,-234.1725 909.9705,-222.82 927.0119,-223.4221 993,-212.1725 1003.5851,-210.368 1014.439,-208.4601 1025.3874,-206.4907"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1026.2474,-209.892 1035.4637,-204.666 1025,-203.004 1026.2474,-209.892"/>
</g>
<!-- t19&#45;&gt;t25 -->
<g id="edge50" class="edge">
<title>t19&#45;&gt;t25</title>
<path fill="none" stroke="#0044a0" d="M342.173,-411.2166C389.8324,-400.6274 451.9326,-386.8297 509.989,-373.9305"/>
<polygon fill="#0044a0" stroke="#0044a0" points="511.0179,-377.2873 520.0207,-371.7016 509.4996,-370.4539 511.0179,-377.2873"/>
<path fill="none" stroke="#0044a0" d="M1502.6864,-411.2619C1471.854,-407.6457 1439.3509,-404.0959 1409,-401.2765 1261.7548,-387.598 1223.4889,-399.5063 1077,-379.2765 1064.9997,-377.6192 1052.6541,-375.6614 1040.2537,-373.5086"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1040.7604,-370.0439 1030.3026,-371.7423 1039.537,-376.9362 1040.7604,-370.0439"/>
</g>
<!-- t20&#45;&gt;t22 -->
<g id="edge33" class="edge">
<title>t20&#45;&gt;t22</title>
<path fill="none" stroke="#0044a0" d="M323.4472,-984.1943C198.6099,-964.2959 37,-926.6746 37,-865.252 37,-865.252 37,-865.252 37,-429.252 37,-211.1917 260.8906,-241.2767 477,-212.1725 490.808,-210.313 505.0108,-208.2741 519.3157,-206.1255"/>
<polygon fill="#0044a0" stroke="#0044a0" points="519.8697,-209.5816 529.2317,-204.6214 518.8199,-202.6608 519.8697,-209.5816"/>
<path fill="none" stroke="#0044a0" d="M1765.9555,-984.2118C1819.0452,-964.5026 1887,-927.1837 1887,-865.2765 1887,-865.2765 1887,-865.2765 1887,-335.7367 1887,-284.8302 1870.8252,-269.3449 1834,-234.197 1824.9302,-225.5403 1814.5141,-217.5334 1803.7754,-210.3065"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1805.4063,-207.1923 1795.1135,-204.6804 1801.5933,-213.0627 1805.4063,-207.1923"/>
</g>
<!-- t26 -->
<g id="node26" class="node">
<title>t26</title>
<polygon fill="#efa03b" stroke="#000000" points="568.0652,-47.332 735,-20.0049 901.9348,-47.332 901.779,-91.5483 568.221,-91.5483 568.0652,-47.332"/>
<polygon fill="none" stroke="#000000" points="564.0538,-43.9306 735,-15.9468 905.9462,-43.9306 905.7643,-95.5532 564.2357,-95.5532 564.0538,-43.9306"/>
<text text-anchor="middle" x="735" y="-53.5531" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.copy_final</text>
<!-- t24 -->
<g id="node24" class="node">
<title>t24</title>
<polygon fill="#efa03b" stroke="#000000" points="1556.0652,-47.332 1723,-20.0049 1889.9348,-47.332 1889.779,-91.5483 1556.221,-91.5483 1556.0652,-47.332"/>
<polygon fill="none" stroke="#000000" points="1552.0538,-43.9306 1723,-15.9468 1893.9462,-43.9306 1893.7643,-95.5532 1552.2357,-95.5532 1552.0538,-43.9306"/>
<text text-anchor="middle" x="1723" y="-53.5531" font-family="Times,serif" font-size="20.00" fill="#000000">ocrmypdf.pipeline.copy_final</text>
</g>
<!-- t22&#45;&gt;t26 -->
<g id="edge49" class="edge">
<title>t22&#45;&gt;t26</title>
<path fill="none" stroke="#0044a0" d="M735,-124.8984C735,-118.6708 735,-112.2475 735,-105.9711"/>
<polygon fill="#0044a0" stroke="#0044a0" points="738.5001,-105.6074 735,-95.6074 731.5001,-105.6075 738.5001,-105.6074"/>
</g>
<!-- t24&#45;&gt;t26 -->
<g id="edge47" class="edge">
<title>t24&#45;&gt;t26</title>
<path fill="none" stroke="#0044a0" d="M368.6395,-140.6178C427.4047,-127.6148 498.1821,-111.9539 561.7955,-97.8781"/>
<polygon fill="#0044a0" stroke="#0044a0" points="562.9329,-101.2112 571.9406,-95.6333 561.4205,-94.3765 562.9329,-101.2112"/>
</g>
<!-- t23&#45;&gt;t26 -->
<g id="edge48" class="edge">
<title>t23&#45;&gt;t26</title>
<path fill="none" stroke="#0044a0" d="M1095.1818,-140.9007C1037.3475,-127.8387 967.4279,-112.0473 904.6388,-97.8663"/>
<polygon fill="#0044a0" stroke="#0044a0" points="905.1511,-94.3939 894.6256,-95.6048 903.6089,-101.2219 905.1511,-94.3939"/>
<!-- t22&#45;&gt;t24 -->
<g id="edge38" class="edge">
<title>t22&#45;&gt;t24</title>
<path fill="none" stroke="#0044a0" d="M1723,-124.9057C1723,-118.6774 1723,-112.2535 1723,-105.9763"/>
<polygon fill="#0044a0" stroke="#0044a0" points="1726.5001,-105.6115 1723,-95.6115 1719.5001,-105.6115 1726.5001,-105.6115"/>
</g>
</g>
</svg>
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 36 KiB

After

Width:  |  Height:  |  Size: 31 KiB

src/ocrmypdf/pipeline.py
+31 -57
View File
@@ -993,13 +993,14 @@ def skip_page(
re_symlink(input_file, output_file, log)
def _merge_pages_common(
def merge_pages(
input_files_groups,
output_file,
log,
context):
"""Determine ordered list of PDF pages to merge. Returns PDF from which
metadata should be drawn if present (for qpdf)."""
options = context.get_options()
input_files = list(f for f in flatten_groups(input_files_groups)
if not f.endswith('.txt'))
@@ -1024,17 +1025,22 @@ def _merge_pages_common(
pdf_pages = sorted(input_files, key=input_file_order)
log.debug("Final pages: " + "\n".join(pdf_pages))
return pdf_pages, metadata_file
args = (pdf_pages, metadata_file, output_file, log, context)
if options.output_type.startswith('pdfa'):
_do_merge_ghostscript(*args)
elif fitz:
_do_merge_mupdf(*args)
else:
_do_merge_qpdf(*args)
def merge_pages_ghostscript(
input_files_groups,
def _do_merge_ghostscript(
pdf_pages,
metadata_file,
output_file,
log,
context):
options = context.get_options()
pdf_pages, _ = _merge_pages_common(
input_files_groups, output_file, log, context)
input_pdfinfo = context.get_pdfinfo()
ghostscript.generate_pdfa(
pdf_version=input_pdfinfo.min_version,
@@ -1052,14 +1058,13 @@ def merge_pages_ghostscript(
os.replace(output_file + '_toc.pdf', output_file)
def merge_pages_qpdf(
input_files_groups,
def _do_merge_qpdf(
pdf_pages,
metadata_file,
output_file,
log,
context):
options = context.get_options()
pdf_pages, metadata_file = _merge_pages_common(
input_files_groups, output_file, log, context)
reader_metadata = pypdf.PdfFileReader(metadata_file)
pdfmark = get_pdfmark(reader_metadata, options)
@@ -1083,18 +1088,15 @@ def merge_pages_qpdf(
log=log)
def merge_pages_mupdf(
input_files_groups,
def _do_merge_mupdf(
pdf_pages,
metadata_file,
output_file,
log,
context):
assert fitz
options = context.get_options()
pdf_pages, metadata_file = _merge_pages_common(
input_files_groups, output_file, log, context)
doc = fitz.Document()
reader_metadata = pypdf.PdfFileReader(metadata_file)
@@ -1346,15 +1348,15 @@ def build_pipeline(options, work_folder, log, context):
task_combine_layers.active_if(options.pdf_renderer == 'hocr' or
options.pdf_renderer == 'sandwich')
task_jbig2_10to1 = main_pipeline.collate(
task_func=jbig2_10to1,
input=[combine_layers],
filter=regex(r".*/(\d{5})(\d)(?:\.rendered\.pdf)"),
output=os.path.join(work_folder, r'\1x.opt.pdf'),
extras=[log, context])
task_jbig2_10to1.graphviz(fillcolor='"#00cc66"')
task_jbig2_10to1.active_if((options.pdf_renderer == 'hocr' or
options.pdf_renderer == 'sandwich'))
# task_jbig2_10to1 = main_pipeline.collate(
# task_func=jbig2_10to1,
# input=[combine_layers],
# filter=regex(r".*/(\d{5})(\d)(?:\.rendered\.pdf)"),
# output=os.path.join(work_folder, r'\1x.opt.pdf'),
# extras=[log, context])
# task_jbig2_10to1.graphviz(fillcolor='"#00cc66"')
# task_jbig2_10to1.active_if((options.pdf_renderer == 'hocr' or
# options.pdf_renderer == 'sandwich'))
# Tesseract OCR+PDF
task_ocr_tesseract_and_render_pdf = main_pipeline.collate(
@@ -1376,7 +1378,6 @@ def build_pipeline(options, work_folder, log, context):
extras=[log, context])
task_generate_postscript_stub.active_if(options.output_type.startswith('pdfa'))
# Bypass valve
task_skip_page = main_pipeline.transform(
task_func=skip_page,
@@ -1387,41 +1388,16 @@ def build_pipeline(options, work_folder, log, context):
extras=[log, context])
# Merge pages
task_merge_pages_ghostscript = main_pipeline.merge(
task_func=merge_pages_ghostscript,
task_merge_pages = main_pipeline.merge(
task_func=merge_pages,
input=[task_combine_layers,
task_repair_and_parse_pdf,
task_render_hocr_debug_page,
task_skip_page,
task_ocr_tesseract_and_render_pdf,
task_generate_postscript_stub],
output=os.path.join(work_folder, 'merged.pdf'),
extras=[log, context])
task_merge_pages_ghostscript.active_if(
options.output_type.startswith('pdfa'))
task_merge_pages_qpdf = main_pipeline.merge(
task_func=merge_pages_qpdf,
input=[task_combine_layers,
task_render_hocr_debug_page,
task_skip_page,
task_ocr_tesseract_and_render_pdf,
task_repair_and_parse_pdf],
output=os.path.join(work_folder, 'merged.pdf'),
extras=[log, context])
task_merge_pages_qpdf.active_if(
options.output_type == 'pdf' and not fitz)
task_merge_pages_mupdf = main_pipeline.merge(
task_func=merge_pages_mupdf,
input=[task_combine_layers,
task_render_hocr_debug_page,
task_skip_page,
task_ocr_tesseract_and_render_pdf,
task_repair_and_parse_pdf],
output=os.path.join(work_folder, 'merged.pdf'),
extras=[log, context])
task_merge_pages_mupdf.active_if(
options.output_type == 'pdf' and fitz)
task_merge_sidecars = main_pipeline.merge(
task_func=merge_sidecars,
@@ -1435,8 +1411,6 @@ def build_pipeline(options, work_folder, log, context):
# Finalize
main_pipeline.merge(
task_func=copy_final,
input=[task_merge_pages_ghostscript,
task_merge_pages_mupdf,
task_merge_pages_qpdf],
input=[task_merge_pages],
output=options.output_file,
extras=[log, context])