diff --git a/README.org b/README.org index 0ebdcb6..65484b1 100644 --- a/README.org +++ b/README.org @@ -49,7 +49,7 @@ PDF=$1 python -m table_ocr.pdf_to_images $PDF | grep .png > /tmp/pdf-images.txt cat /tmp/pdf-images.txt | xargs -I{} python -m table_ocr.extract_tables {} | grep table > /tmp/extracted-tables.txt cat /tmp/extracted-tables.txt | xargs -I{} python -m table_ocr.extract_cells {} | grep cells > /tmp/extracted-cells.txt -cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {} --psm 7 -l table-ocr +cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {} for image in $(cat /tmp/extracted-tables.txt); do dir=$(dirname $image) @@ -57,6 +57,7 @@ for image in $(cat /tmp/extracted-tables.txt); do done #+END_SRC + The package was written in a [[https://en.wikipedia.org/wiki/Literate_programming][literate programming]] style. The source code at [[https://eihli.github.io/image-table-ocr/pdf_table_extraction_and_ocr.html]] is meant to act as the documentation and reference material. diff --git a/ocr_tables b/ocr_tables index 25e936a..5f2e413 100755 --- a/ocr_tables +++ b/ocr_tables @@ -5,8 +5,7 @@ PDF=$1 python -m table_ocr.pdf_to_images $PDF | grep .png > /tmp/pdf-images.txt cat /tmp/pdf-images.txt | xargs -I{} python -m table_ocr.extract_tables {} | grep table > /tmp/extracted-tables.txt cat /tmp/extracted-tables.txt | xargs -I{} python -m table_ocr.extract_cells {} | grep cells > /tmp/extracted-cells.txt -cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {} --psm 7 -l table-ocr - +cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {} for image in $(cat /tmp/extracted-tables.txt); do dir=$(dirname $image) python -m table_ocr.ocr_to_csv $(find $dir/cells -name "*.txt") diff --git a/pdf_table_extraction_and_ocr.html b/pdf_table_extraction_and_ocr.html index be38955..5126f4d 100644 --- a/pdf_table_extraction_and_ocr.html +++ b/pdf_table_extraction_and_ocr.html @@ -3,7 +3,7 @@ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> - + PDF Parsing @@ -225,94 +225,98 @@

Table of Contents

-
-

1 Overview

+
+

1 Overview

This Python package provides utilities for extracting tabular data from PDF @@ -359,17 +363,17 @@ The package is split into modules with narrow focuses.

  • pdf_to_images uses Poppler and ImageMagick to extract images from a PDF.
  • extract_tables finds and extracts table-looking things from an image.
  • extract_cells extracts and orders cells from a table.
    • -
  • ocr_image uses Tesseract to turn a OCR the text from an image of a cell.
    • +
  • ocr_image uses Tesseract to OCR the text from an image of a cell.
  • ocr_to_csv converts into a CSV the directory structure that ocr_image outputs.
    • -
    -

    1.1 Requirements

    +
    +

    1.1 Requirements

    -
    -

    1.1.1 Python packages

    +
    +

    1.1.1 Python packages

    • numpy
      • @@ -379,8 +383,8 @@ The package is split into modules with narrow focuses.
    -
    -

    1.1.2 External

    +
    +

    1.1.2 External

    • pdfimages from Poppler
      • @@ -391,8 +395,8 @@ The package is split into modules with narrow focuses.
    -
    -

    1.2 Contributing

    +
    +

    1.2 Contributing

    This package was created in a literate programming style with the help of Babel. @@ -405,8 +409,8 @@ barrier for contributors who aren’t already familiar with Emacs and Babel.

    -
    -

    1.3 Example usage

    +
    +

    1.3 Example usage

    Here is an example of a shell script that uses each module to turn a pdf with a @@ -424,26 +428,29 @@ you need into your own python projects and use them as needed.

    -
    #!/bin/sh
+
    #!/bin/sh
 
 PDF=$1
 
 python -m table_ocr.pdf_to_images $PDF | grep .png > /tmp/pdf-images.txt
 cat /tmp/pdf-images.txt | xargs -I{} python -m table_ocr.extract_tables {}  | grep table > /tmp/extracted-tables.txt
 cat /tmp/extracted-tables.txt | xargs -I{} python -m table_ocr.extract_cells {} | grep cells > /tmp/extracted-cells.txt
-cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {} --psm 7 -l table-ocr
-
+cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {}
 for image in $(cat /tmp/extracted-tables.txt); do
     dir=$(dirname $image)
     python -m table_ocr.ocr_to_csv $(find $dir/cells -name "*.txt")
 done
    + +

    +Any extra args you pass after the image path to python -m table_ocr.ocr_image will be passed directly to tesseract as options. If you don’t pass anything, reasonable english defaults are used. +

    -
    -

    1.4 Possible improvements

    +
    +

    1.4 Possible improvements

    Detect text with the stroke-width-transform alogoritm. https://zablo.net/blog/post/stroke-width-transform-swt-python/index.html @@ -452,8 +459,8 @@ Detect text with the stroke-width-transform alogoritm. -

    2 Preparing data

    +
    +

    2 Preparing data

    Not all pdfs need to be sent through OCR to extract the text content. If you can @@ -462,27 +469,27 @@ probably aren’t necessary.

    -
    -

    2.1 Converting PDFs to images

    +
    +

    2.1 Converting PDFs to images

    This code calls out to pdfimages from Poppler.

    -
    # Wrapper around the Poppler command line utility "pdfimages" and helpers for
+
    # Wrapper around the Poppler command line utility "pdfimages" and helpers for
 # finding the output files of that command.
 def pdf_to_images(pdf_filepath):
     """
     Turn a pdf into images
+    Returns the filenames of the created images sorted lexicographically.
     """
     directory, filename = os.path.split(pdf_filepath)
-    with working_dir(directory):
-        image_filenames = pdfimages(pdf_filepath)
+    image_filenames = pdfimages(pdf_filepath)
 
     # Since pdfimages creates a number of files named each for there page number
     # and doesn't return us the list that it created
-    return [os.path.join(directory, f) for f in image_filenames]
+    return sorted([os.path.join(directory, f) for f in image_filenames])
 
 
 def pdfimages(pdf_filepath):
@@ -495,8 +502,14 @@ This code calls out to     uses 3 digits in its regex.
     """
     directory, filename = os.path.split(pdf_filepath)
+    if not os.path.isabs(directory):
+        directory = os.path.abspath(directory)
     filename_sans_ext = filename.split(".pdf")[0]
-    subprocess.run(["pdfimages", "-png", pdf_filepath, filename.split(".pdf")[0]])
+
+    # pdfimages outputs results to the current working directory
+    with working_dir(directory):
+        subprocess.run(["pdfimages", "-png", filename, filename.split(".pdf")[0]])
+
     image_filenames = find_matching_files_in_dir(filename_sans_ext, directory)
     logger.debug(
         "Converted {} into files:\n{}".format(pdf_filepath, "\n".join(image_filenames))
@@ -516,8 +529,8 @@ This code calls out to 
-
    2.2 Detecting image orientation and applying rotation.
    
+
    
+
    2.2 Detecting image orientation and applying rotation.
    
 
    
 
    
 Tesseract can detect orientation and we can then use ImageMagick’s mogrify to
@@ -546,19 +559,29 @@ to correct the rotation. This makes OCR more straightforward.
 
    
 
 
    
-
    def preprocess_img(filepath):
-    """
-    Processing that involves running shell executables,
+
    def preprocess_img(filepath, tess_params=None):
+    """Processing that involves running shell executables,
     like mogrify to rotate.
+
+    Uses tesseract to detect rotation.
+   
+    Orientation and script detection is only available for legacy tesseract
+    (--oem 0). Some versions of tesseract will segfault if you let it run OSD
+    with the default oem (3).
     """
-    rotate = get_rotate(filepath)
+    if tess_params is None:
+        tess_params = ["--psm", "0", "--oem", "0"]
+    rotate = get_rotate(filepath, tess_params)
     logger.debug("Rotating {} by {}.".format(filepath, rotate))
     mogrify(filepath, rotate)
 
 
-def get_rotate(image_filepath):
+def get_rotate(image_filepath, tess_params):
+    """
+    """
+    tess_command = ["tesseract"] + tess_params + [image_filepath, "-"]
     output = (
-        subprocess.check_output(["tesseract", "--psm", "0", image_filepath, "-"])
+        subprocess.check_output(tess_command)
         .decode("utf-8")
         .split("\n")
     )
@@ -575,8 +598,8 @@ to correct the rotation. This makes OCR more straightforward.
 
    
 
    
 
-
    
-
    3 Detecting tables
    
+
    
+
    3 Detecting tables
    
 
    
 
    
 This answer from opencv.org was heavily referenced while writing the code around
@@ -597,7 +620,7 @@ that makes things like shape detection more accurate.
 
    
 
 
    
-
    def find_tables(image):
+
    def find_tables(image):
     BLUR_KERNEL_SIZE = (17, 17)
     STD_DEV_X_DIRECTION = 0
     STD_DEV_Y_DIRECTION = 0
@@ -680,8 +703,8 @@ cv2.imwrite("resources/examples/example-table.png"
 
    
 
    
 
-
    
-
    3.1 Improving accuracy
    
+
    
+
    3.1 Improving accuracy
    
 
    
 
    
 It’s likely that some images will contain tables that aren’t accurately
@@ -702,8 +725,8 @@ the x, y, width and height are anywhere in that range.
 
    
 
    
 
-
    
-
    4 OCR tables
    
+
    
+
    4 OCR tables
    
 
    
 
    
 Tesseract does not perform well when run on images of tables. It performs best
@@ -720,9 +743,13 @@ We’ll start with an image shown at the end of the previous section.
 
    
 
    
 
-
    
-
    4.1 Training Tesseract
    
+
    
+
    4.1 Training Tesseract
    
 
    
+
    
+Tesseract is used for recognizing characters. It is not involved in extracting the tables from an image or in extracting cells from the table.
+
    
+
 
    
 It’s a very good idea to train tesseract. Accuracy will improve tremendously.
 
    
@@ -732,7 +759,7 @@ Clone the tesstrain repo at 
 
 
    
-Run the ocr_tables script on a few pdfs to generate some training data. That
+Run the ocr_tables script on a few pdfs to generate some training data. That
 script outputs pairs of .png and .gt.txt files that can be used by
 tesstrain.
 
    
@@ -772,10 +799,70 @@ Once the training is complete, there will be a new file
 Tesseract searches for models. On my machine, it was /usr/local/share/tessdata/.
 
    
 
    
+
+
    
+
    4.1.1 Training tips
    
+
    
+
    
+Here is a tip for quickly creating training data.
+
    
+
+
    
+The output of the ocr_cells script will be a directory named ocr_data that
+will have two files for each cell. One file is the image of the cell and the
+other file is the OCR text.
+
    
+
+
    
+You’ll want to compare each image to its OCR text to check for accuracy. If
+the text doesn’t match, you’ll want to update the text and add the image to the
+training data.
+
    
+
+
    
+The fastest way to do this is with feh.
+
    
+
+
    
+feh lets you view an image and a caption at the same time and lets you edit
+the caption from within feh.
+
    
+
+
    
+feh expects the captions to be named <image-name>.txt, so use a little
+shell-fu to do a quick rename.
+
    
+
+
    
+
    for f in *.txt; do f1=$(cut -d"." -f1 <(echo $f)); mv $f ${f1}.png.txt; done
+
    
 
    
 
-
    
-
    4.2 Blur
    
+
    
+Then run feh -K . to specify the current directory as the caption directory.
+This will open a window with the first image in the directory and its caption.
+
    
+
+
    
+Press c to edit the caption (if needed) and n~/~p to move to the
+next/previons images. Press q to quit.
+
    
+
+
    
+When finished, rename the files back to the filename structure that Tesseract
+looks for in training.
+
    
+
+
    
+
    for f in *.txt; do f1=$(cut -d"." -f1 <(echo $f)); mv $f ${f1}.gt.txt; done
+
    
+
    
+
    
+
    
+
    
+
+
    
+
    4.2 Blur
    
 
    
 
    
 Blurring helps to make noise less noisy so that the overall structure of an
@@ -815,8 +902,8 @@ cv2.imwrite("resources/examples/example-table-blur
 
    
 
    
 
-
    
-
    4.3 Threshold
    
+
    
+
    4.3 Threshold
    
 
    
 
    
 We’ve got a bunch of pixels that are gray. Thresholding will turn them all
@@ -854,8 +941,8 @@ cv2.imwrite("resources/examples/example-table-thre
 
    
 
    
 
-
    
-
    4.4 Finding the vertical and horizontal lines of the table
    
+
    
+
    4.4 Finding the vertical and horizontal lines of the table
    
 
    
 
    
 
    vertical = horizontal = img_bin.copy()
@@ -894,8 +981,8 @@ cv2.imwrite("resources/examples/example-table-line
 
    
 
    
 
-
    
-
    4.5 Finding the contours
    
+
    
+
    4.5 Finding the contours
    
 
    
 
    
 Blurring and thresholding allow us to find the lines. Opening the lines allows
@@ -944,7 +1031,7 @@ above/below certain sizes.
 
    
 
 
    
-
    contours, heirarchy = cv2.findContours(
+
    contours, heirarchy = cv2.findContours(
     mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE,
 )
 
@@ -976,8 +1063,8 @@ above/below certain sizes.
 
    
 
    
 
-
    
-
    4.6 Sorting the bounding rectangles
    
+
    
+
    4.6 Sorting the bounding rectangles
    
 
    
 
    
 We want to process these from left-to-right, top-to-bottom.
@@ -996,7 +1083,7 @@ value of their center. We’ll remove those rectangles from the list and rep
 
    
 
 
    
-
    def cell_in_same_row(c1, c2):
+
    def cell_in_same_row(c1, c2):
     c1_center = c1[1] + c1[3] - c1[3] / 2
     c2_bottom = c2[1] + c2[3]
     c2_top = c2[1]
@@ -1070,7 +1157,7 @@ cv2.imwrite("resources/examples/example-table-cell
 
    
 
 
    
-
    def extract_cell_images_from_table(image):
+
    def extract_cell_images_from_table(image):
     BLUR_KERNEL_SIZE = (17, 17)
     STD_DEV_X_DIRECTION = 0
     STD_DEV_Y_DIRECTION = 0
@@ -1184,8 +1271,8 @@ cv2.imwrite("resources/examples/example-table-cell
 
    
 
    
 
-
    
-
    4.7 Cropping each cell to the text
    
+
    
+
    4.7 Cropping each cell to the text
    
 
    
 
    
 OCR with Tesseract works best when there is about 10 pixels of white border
@@ -1272,8 +1359,8 @@ cv2.imwrite("resources/examples/example-table-cell
 
    
 
    
 
-
    
-
    4.8 OCR each cell
    
+
    
+
    4.8 OCR each cell
    
 
    
 
    
 If we cleaned up the images well enough, we might get some accurate OCR!
@@ -1303,10 +1390,31 @@ period into a comma, then you might need to do some custom Tesseract training.
    +

    +The second argument passed to ocr_image is a string of the command line arguments passed directly to tesseract. You can view the available options at https://github.com/tesseract-ocr/tesseract/blob/master/doc/tesseract.1.asc#options +

    + +

    +If no options are passed to tesseract, then language defaults to english. This means tesseract needs to be able to find a file named eng.traineddata on whatever path it searches for languages. +

    + +

    +This python package comes with eng.traineddata and table-ocr.traineddata. table-ocr.traineddata is a personal model that I’ve found to be more accurate for my use case. You should train your own to maximize accuracy. +

    + +

    +When you pip install this package, the traineddata gets copied to a tessdata folder in the same directory in which pip installs the package. +

    + +

    +The ocr_image package in this repo defaults to using the --tessdata-dir option to the package’s tessdata directory in the package install location and the -l option to the table_ocr language. +

    + 
    import pytesseract
 import cv2
 import numpy as np
+import math
 image = cv2.imread("resources/examples/example-table-cell-1-1.png", cv2.IMREAD_GRAYSCALE)
 <<crop-to-text>>
 <<ocr-image>>
@@ -1322,8 +1430,8 @@ ocr_image(image, "--psm 7")
    -
    -

    5 Files

    +
    +

    5 Files

    @@ -1331,8 +1439,8 @@ ocr_image(image, "--psm 7")
    -
    -

    5.1 setup.py

    +
    +

    5.1 setup.py

    import setuptools
@@ -1340,43 +1448,43 @@ ocr_image(image, "--psm 7")
 long_description = """
 Utilities for turning images of tables into CSV data. Uses Tesseract and OpenCV.
 
-Requires binaries for tesseract and pdfimages (from Poppler).
+Requires binaries for tesseract, ImageMagick, and pdfimages (from Poppler).
 """
 setuptools.setup(
     name="table_ocr",
-    version="0.0.1",
+    version="0.2.0",
     author="Eric Ihli",
     author_email="eihli@owoga.com",
-    description="Turn images of tables into CSV data.",
+    description="Extract text from tables in images.",
     long_description=long_description,
     long_description_content_type="text/plain",
     url="https://github.com/eihli/image-table-ocr",
     packages=setuptools.find_packages(),
+    package_data={
+        "table_ocr": ["tessdata/table-ocr.traineddata", "tessdata/eng.traineddata"]
+    },
     classifiers=[
         "Programming Language :: Python :: 3",
         "License :: OSI Approved :: MIT License",
         "Operating System :: OS Independent",
     ],
-    install_requires=[
-        "pytesseract~=0.3",
-        "opencv-python~=4.2",
-    ],
-    python_requires='>=3.6',
+    install_requires=["pytesseract~=0.3", "opencv-python~=4.2",],
+    python_requires=">=3.6",
 )
    -
    -

    5.2 table_ocr

    +
    +

    5.2 table_ocr

    -
    -

    5.2.1 table_ocr/__init__.py

    +
    +

    5.2.1 table_ocr/__init__.py

    -
    -

    5.2.2 table_ocr/util.py

    +
    +

    5.2.2 table_ocr/util.py

    from contextlib import contextmanager
@@ -1413,15 +1521,15 @@ setuptools.setup(
    -
    -

    5.2.3 table_ocr/pdf_to_images/

    +
    +

    5.2.3 table_ocr/pdf_to_images/

    -
    -
    5.2.3.1 table_ocr/pdf_to_images/__init__.py
    +
    +
    5.2.3.1 table_ocr/pdf_to_images/__init__.py
    -
    import os
+
    import os
 import re
 import subprocess
 
@@ -1434,14 +1542,14 @@ setuptools.setup(
 def pdf_to_images(pdf_filepath):
     """
     Turn a pdf into images
+    Returns the filenames of the created images sorted lexicographically.
     """
     directory, filename = os.path.split(pdf_filepath)
-    with working_dir(directory):
-        image_filenames = pdfimages(pdf_filepath)
+    image_filenames = pdfimages(pdf_filepath)
 
     # Since pdfimages creates a number of files named each for there page number
     # and doesn't return us the list that it created
-    return [os.path.join(directory, f) for f in image_filenames]
+    return sorted([os.path.join(directory, f) for f in image_filenames])
 
 
 def pdfimages(pdf_filepath):
@@ -1454,8 +1562,14 @@ setuptools.setup(
     uses 3 digits in its regex.
     """
     directory, filename = os.path.split(pdf_filepath)
+    if not os.path.isabs(directory):
+        directory = os.path.abspath(directory)
     filename_sans_ext = filename.split(".pdf")[0]
-    subprocess.run(["pdfimages", "-png", pdf_filepath, filename.split(".pdf")[0]])
+
+    # pdfimages outputs results to the current working directory
+    with working_dir(directory):
+        subprocess.run(["pdfimages", "-png", filename, filename.split(".pdf")[0]])
+
     image_filenames = find_matching_files_in_dir(filename_sans_ext, directory)
     logger.debug(
         "Converted {} into files:\n{}".format(pdf_filepath, "\n".join(image_filenames))
@@ -1471,19 +1585,29 @@ setuptools.setup(
     ]
     return files
 
-def preprocess_img(filepath):
-    """
-    Processing that involves running shell executables,
+def preprocess_img(filepath, tess_params=None):
+    """Processing that involves running shell executables,
     like mogrify to rotate.
+
+    Uses tesseract to detect rotation.
+   
+    Orientation and script detection is only available for legacy tesseract
+    (--oem 0). Some versions of tesseract will segfault if you let it run OSD
+    with the default oem (3).
     """
-    rotate = get_rotate(filepath)
+    if tess_params is None:
+        tess_params = ["--psm", "0", "--oem", "0"]
+    rotate = get_rotate(filepath, tess_params)
     logger.debug("Rotating {} by {}.".format(filepath, rotate))
     mogrify(filepath, rotate)
 
 
-def get_rotate(image_filepath):
+def get_rotate(image_filepath, tess_params):
+    """
+    """
+    tess_command = ["tesseract"] + tess_params + [image_filepath, "-"]
     output = (
-        subprocess.check_output(["tesseract", "--psm", "0", image_filepath, "-"])
+        subprocess.check_output(tess_command)
         .decode("utf-8")
         .split("\n")
     )
@@ -1499,8 +1623,8 @@ setuptools.setup(
    -
    -
    5.2.3.2 table_ocr/pdf_to_images/__main__.py
    +
    +
    5.2.3.2 table_ocr/pdf_to_images/__main__.py

    Takes a variable number of pdf files and creates images out of each page of the @@ -1520,7 +1644,7 @@ blank line.

    -
    import argparse
+
    import argparse
 
 from table_ocr.util import working_dir, make_tempdir, get_logger
 from table_ocr.pdf_to_images import pdf_to_images, preprocess_img
@@ -1553,15 +1677,16 @@ parser.add_argument("files", nargs=
    -
    -

    5.2.4 table_ocr/extract_tables/

    +
    +

    5.2.4 table_ocr/extract_tables/

    -
    -
    5.2.4.1 table_ocr/extract_tables/__init__.py
    +
    +
    5.2.4.1 table_ocr/extract_tables/__init__.py
    -
    import cv2
+
    import os
+import cv2
 
 def find_tables(image):
     BLUR_KERNEL_SIZE = (17, 17)
@@ -1610,13 +1735,35 @@ parser.add_argument("files", nargs=    # Leaving that step as a future TODO if it is ever necessary.
     images = [image[y:y+h, x:x+w] for x, y, w, h in bounding_rects]
     return images
+
+def main(files):
+    results = []
+    for f in files:
+        directory, filename = os.path.split(f)
+        image = cv2.imread(f, cv2.IMREAD_GRAYSCALE)
+        tables = find_tables(image)
+        files = []
+        filename_sans_extension = os.path.splitext(filename)[0]
+        if tables:
+            os.makedirs(os.path.join(directory, filename_sans_extension), exist_ok=True)
+        for i, table in enumerate(tables):
+            table_filename = "table-{:03d}.png".format(i)
+            table_filepath = os.path.join(
+                directory, filename_sans_extension, table_filename
+            )
+            files.append(table_filepath)
+            cv2.imwrite(table_filepath, table)
+        if tables:
+            results.append((f, files))
+    # Results is [[<input image>, [<images of detected tables>]]]
+    return results
    -
    -
    5.2.4.2 table_ocr/extract_tables/__main__.py
    +
    +
    5.2.4.2 table_ocr/extract_tables/__main__.py

    Takes 1 or more image paths as arguments. @@ -1648,60 +1795,33 @@ For each image path given as an agument, outputs:

    -
    import argparse
-import os
-
-import cv2
+
    import argparse
 
-from table_ocr.extract_tables import find_tables
+from table_ocr.extract_tables import main
 
 parser = argparse.ArgumentParser()
 parser.add_argument("files", nargs="+")
-
-
-def main(files):
-    results = []
-    for f in files:
-        directory, filename = os.path.split(f)
-        image = cv2.imread(f, cv2.IMREAD_GRAYSCALE)
-        tables = find_tables(image)
-        files = []
-        filename_sans_extension = os.path.splitext(filename)[0]
-        if tables:
-            os.makedirs(os.path.join(directory, filename_sans_extension), exist_ok=True)
-        for i, table in enumerate(tables):
-            table_filename = "table-{:03d}.png".format(i)
-            table_filepath = os.path.join(
-                directory, filename_sans_extension, table_filename
-            )
-            files.append(table_filepath)
-            cv2.imwrite(table_filepath, table)
-        if tables:
-            results.append((f, files))
-
-    for image_filename, table_filenames in results:
-        print("\n".join(table_filenames))
-
-
-if __name__ == "__main__":
-    args = parser.parse_args()
-    files = args.files
-    main(files)
+args = parser.parse_args()
+files = args.files
+results = main(files)
+for image, tables in results:
+    print("\n".join(tables))
    -
    -

    5.2.5 table_ocr/extract_cells/

    +
    +

    5.2.5 table_ocr/extract_cells/

    -
    -
    5.2.5.1 table_ocr/extract_cells/__init__.py
    +
    +
    5.2.5.1 table_ocr/extract_cells/__init__.py
    import cv2
+import os
 
 def extract_cell_images_from_table(image):
     BLUR_KERNEL_SIZE = (17, 17)
@@ -1798,13 +1918,29 @@ parser.add_argument("files", nargs=            cell_images_row.append(image[y:y+h, x:x+w])
         cell_images_rows.append(cell_images_row)
     return cell_images_rows
+
+def main(f):
+    results = []
+    directory, filename = os.path.split(f)
+    table = cv2.imread(f, cv2.IMREAD_GRAYSCALE)
+    rows = extract_cell_images_from_table(table)
+    cell_img_dir = os.path.join(directory, "cells")
+    os.makedirs(cell_img_dir, exist_ok=True)
+    paths = []
+    for i, row in enumerate(rows):
+        for j, cell in enumerate(row):
+            cell_filename = "{:03d}-{:03d}.png".format(i, j)
+            path = os.path.join(cell_img_dir, cell_filename)
+            cv2.imwrite(path, cell)
+            paths.append(path)
+    return paths
    -
    -
    5.2.5.2 table_ocr/extract_cells/__main__.py
    +
    +
    5.2.5.2 table_ocr/extract_cells/__main__.py

    Takes as a command line argument a path to an image of a table. @@ -1827,146 +1963,61 @@ cells.

    -
    import os
-import sys
-
-import cv2
-
-from table_ocr.extract_cells import extract_cell_images_from_table
+
    import sys
 
-def main(f):
-    results = []
-    directory, filename = os.path.split(f)
-    table = cv2.imread(f, cv2.IMREAD_GRAYSCALE)
-    rows = extract_cell_images_from_table(table)
-    cell_img_dir = os.path.join(directory, "cells")
-    os.makedirs(cell_img_dir, exist_ok=True)
-    for i, row in enumerate(rows):
-        for j, cell in enumerate(row):
-            cell_filename = "{:03d}-{:03d}.png".format(i, j)
-            path = os.path.join(cell_img_dir, cell_filename)
-            cv2.imwrite(path, cell)
-            print(path)
-
-
-def extract_cell_images_from_table(image):
-    BLUR_KERNEL_SIZE = (17, 17)
-    STD_DEV_X_DIRECTION = 0
-    STD_DEV_Y_DIRECTION = 0
-    blurred = cv2.GaussianBlur(image, BLUR_KERNEL_SIZE, STD_DEV_X_DIRECTION, STD_DEV_Y_DIRECTION)
-    MAX_COLOR_VAL = 255
-    BLOCK_SIZE = 15
-    SUBTRACT_FROM_MEAN = -2
-    
-    img_bin = cv2.adaptiveThreshold(
-        ~blurred,
-        MAX_COLOR_VAL,
-        cv2.ADAPTIVE_THRESH_MEAN_C,
-        cv2.THRESH_BINARY,
-        BLOCK_SIZE,
-        SUBTRACT_FROM_MEAN,
-    )
-    vertical = horizontal = img_bin.copy()
-    SCALE = 5
-    image_width, image_height = horizontal.shape
-    horizontal_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (int(image_width / SCALE), 1))
-    horizontally_opened = cv2.morphologyEx(img_bin, cv2.MORPH_OPEN, horizontal_kernel)
-    vertical_kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (1, int(image_height / SCALE)))
-    vertically_opened = cv2.morphologyEx(img_bin, cv2.MORPH_OPEN, vertical_kernel)
-    
-    horizontally_dilated = cv2.dilate(horizontally_opened, cv2.getStructuringElement(cv2.MORPH_RECT, (40, 1)))
-    vertically_dilated = cv2.dilate(vertically_opened, cv2.getStructuringElement(cv2.MORPH_RECT, (1, 60)))
-    
-    mask = horizontally_dilated + vertically_dilated
-    contours, heirarchy = cv2.findContours(
-        mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE,
-    )
-    
-    perimeter_lengths = [cv2.arcLength(c, True) for c in contours]
-    epsilons = [0.05 * p for p in perimeter_lengths]
-    approx_polys = [cv2.approxPolyDP(c, e, True) for c, e in zip(contours, epsilons)]
-    
-    # Filter out contours that aren't rectangular. Those that aren't rectangular
-    # are probably noise.
-    approx_rects = [p for p in approx_polys if len(p) == 4]
-    bounding_rects = [cv2.boundingRect(a) for a in approx_polys]
-    
-    # Filter out rectangles that are too narrow or too short.
-    MIN_RECT_WIDTH = 40
-    MIN_RECT_HEIGHT = 10
-    bounding_rects = [
-        r for r in bounding_rects if MIN_RECT_WIDTH < r[2] and MIN_RECT_HEIGHT < r[3]
-    ]
-    
-    # The largest bounding rectangle is assumed to be the entire table.
-    # Remove it from the list. We don't want to accidentally try to OCR
-    # the entire table.
-    largest_rect = max(bounding_rects, key=lambda r: r[2] * r[3])
-    bounding_rects = [b for b in bounding_rects if b is not largest_rect]
-    
-    cells = [c for c in bounding_rects]
-    def cell_in_same_row(c1, c2):
-        c1_center = c1[1] + c1[3] - c1[3] / 2
-        c2_bottom = c2[1] + c2[3]
-        c2_top = c2[1]
-        return c2_top < c1_center < c2_bottom
-    
-    orig_cells = [c for c in cells]
-    rows = []
-    while cells:
-        first = cells[0]
-        rest = cells[1:]
-        cells_in_same_row = sorted(
-            [
-                c for c in rest
-                if cell_in_same_row(c, first)
-            ],
-            key=lambda c: c[0]
-        )
-    
-        row_cells = sorted([first] + cells_in_same_row, key=lambda c: c[0])
-        rows.append(row_cells)
-        cells = [
-            c for c in rest
-            if not cell_in_same_row(c, first)
-        ]
-    
-    # Sort rows by average height of their center.
-    def avg_height_of_center(row):
-        centers = [y + h - h / 2 for x, y, w, h in row]
-        return sum(centers) / len(centers)
-    
-    rows.sort(key=avg_height_of_center)
-    cell_images_rows = []
-    for row in rows:
-        cell_images_row = []
-        for x, y, w, h in row:
-            cell_images_row.append(image[y:y+h, x:x+w])
-        cell_images_rows.append(cell_images_row)
-    return cell_images_rows
+from table_ocr.extract_cells import main
 
-if __name__ == "__main__":
-    main(sys.argv[1])
+paths = main(sys.argv[1])
+print("\n".join(paths))
    -
    -

    5.2.6 table_ocr/ocr_image/

    +
    +

    5.2.6 table_ocr/ocr_image/

    -
    -
    5.2.6.1 table_ocr/ocr_image/__init__.py
    +
    +
    5.2.6.1 table_ocr/ocr_image/__init__.py
    import math
+import os
+import sys
 
 import cv2
 import numpy as np
 import pytesseract
 
+def main(image_file, tess_args):
+    """
+    OCR the image and output the text to a file with an extension that is ready
+    to be used in Tesseract training (.gt.txt).
+
+    Tries to crop the image so that only the relevant text gets passed to Tesseract.
+
+    Returns the name of the text file that contains the text.
+    """
+    directory, filename = os.path.split(image_file)
+    filename_sans_ext, ext = os.path.splitext(filename)
+    image = cv2.imread(image_file, cv2.IMREAD_GRAYSCALE)
+    cropped = crop_to_text(image)
+    ocr_data_dir = os.path.join(directory, "ocr_data")
+    os.makedirs(ocr_data_dir, exist_ok=True)
+    out_imagepath = os.path.join(ocr_data_dir, filename)
+    out_txtpath = os.path.join(ocr_data_dir, "{}.gt.txt".format(filename_sans_ext))
+    cv2.imwrite(out_imagepath, cropped)
+    if not tess_args:
+        d = os.path.dirname(sys.modules["table_ocr"].__file__)
+        tessdata_dir = os.path.join(d, "tessdata")
+        tess_args = ["--psm", "7", "-l", "table-ocr", "--tessdata-dir", tessdata_dir]
+    txt = ocr_image(cropped, " ".join(tess_args))
+    with open(out_txtpath, "w") as txt_file:
+        txt_file.write(txt)
+    return out_txtpath
+
 def crop_to_text(image):
     MAX_COLOR_VAL = 255
     BLOCK_SIZE = 15
@@ -2022,8 +2073,8 @@ cells.
    -
    -
    5.2.6.2 table_ocr/ocr_image/__main__.py
    +
    +
    5.2.6.2 table_ocr/ocr_image/__main__.py

    This does a little bit of cleanup before sending it through tesseract. @@ -2036,13 +2087,8 @@ Creates images and text files that can be used for training tesseract. See 

    import argparse
-import math
-import os
-import sys
-
-import cv2
 
-from table_ocr.ocr_image import crop_to_text, ocr_image
+from table_ocr.ocr_image import main
 
 description="""Takes a single argument that is the image to OCR.
 Remaining arguments are passed directly to Tesseract.
@@ -2053,35 +2099,19 @@ Creates images and text files that can be used for training tesseract. See
 parser = argparse.ArgumentParser(description=description)
 parser.add_argument("image", help="filepath of image to perform OCR")
 
-def main(image_file, tess_args):
-    directory, filename = os.path.split(image_file)
-    filename_sans_ext, ext = os.path.splitext(filename)
-    image = cv2.imread(image_file, cv2.IMREAD_GRAYSCALE)
-    cropped = crop_to_text(image)
-    ocr_data_dir = os.path.join(directory, "ocr_data")
-    os.makedirs(ocr_data_dir, exist_ok=True)
-    out_imagepath = os.path.join(ocr_data_dir, filename)
-    out_txtpath = os.path.join(ocr_data_dir, "{}.gt.txt".format(filename_sans_ext))
-    cv2.imwrite(out_imagepath, cropped)
-    txt = ocr_image(cropped, " ".join(tess_args))
-    print(txt)
-    with open(out_txtpath, "w") as txt_file:
-        txt_file.write(txt)
-
-if __name__ == "__main__":
-    args, tess_args = parser.parse_known_args()
-    main(args.image, tess_args)
+args, tess_args = parser.parse_known_args()
+print(main(args.image, tess_args))
    -
    -

    5.2.7 table_ocr/ocr_to_csv/

    +
    +

    5.2.7 table_ocr/ocr_to_csv/

    -
    -
    5.2.7.1 table_ocr/ocr_to_csv/__init__.py
    +
    +
    5.2.7.1 table_ocr/ocr_to_csv/__init__.py
    import csv
@@ -2115,8 +2145,8 @@ parser.add_argument("image", 
-
    5.2.7.2 table_ocr/ocr_to_csv/__main__.py
    
+
    
+
    5.2.7.2 table_ocr/ocr_to_csv/__main__.py
    
 
    
 
    
 
    import argparse
@@ -2145,8 +2175,8 @@ parser.add_argument("files", nargs=
 
    
 
-
    
-
    6 Utils
    
+
    
+
    6 Utils
    
 
    
 
    
 The following code lets us specify a size for images when they are exported to
@@ -2173,7 +2203,7 @@ with advice-add.
 
    
 
 
    
-
    (concat "#+ATTR_HTML: :width " width " :height " height "\n[[file:" text "]]")
+
    (concat "#+ATTR_HTML: :width " width " :height " height "\n[[file:" text "]]")
 
    
 
    
 
@@ -2195,8 +2225,8 @@ with advice-add.
 
    
 
    
 
-
    
-
    6.1 Logging
    
+
    
+
    6.1 Logging
    
 
    
 
    
 
    def get_logger(name):
@@ -2217,7 +2247,7 @@ with advice-add.
 
    
 
    
 
    Author: Eric Ihli
    
-
    Created: 2020-04-25 Sat 12:20
    
+
    Created: 2020-10-14 Wed 21:28
    
 
    
 
 
diff --git a/pdf_table_extraction_and_ocr.org b/pdf_table_extraction_and_ocr.org
index 3d1587b..bc86c47 100644
--- a/pdf_table_extraction_and_ocr.org
+++ b/pdf_table_extraction_and_ocr.org
@@ -97,14 +97,15 @@ PDF=$1
 python -m table_ocr.pdf_to_images $PDF | grep .png > /tmp/pdf-images.txt
 cat /tmp/pdf-images.txt | xargs -I{} python -m table_ocr.extract_tables {}  | grep table > /tmp/extracted-tables.txt
 cat /tmp/extracted-tables.txt | xargs -I{} python -m table_ocr.extract_cells {} | grep cells > /tmp/extracted-cells.txt
-cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {} --psm 7 -l table-ocr
-
+cat /tmp/extracted-cells.txt | xargs -I{} python -m table_ocr.ocr_image {}
 for image in $(cat /tmp/extracted-tables.txt); do
     dir=$(dirname $image)
     python -m table_ocr.ocr_to_csv $(find $dir/cells -name "*.txt")
 done
 #+END_SRC
 
+Any extra args you pass after the image path to ~python -m table_ocr.ocr_image~ will be passed directly to tesseract as options. If you don't pass anything, reasonable english defaults are used.
+
 ** Possible improvements
 
 Detect text with the stroke-width-transform alogoritm. https://zablo.net/blog/post/stroke-width-transform-swt-python/index.html
@@ -199,7 +200,7 @@ def preprocess_img(filepath, tess_params=None):
     like mogrify to rotate.
 
     Uses tesseract to detect rotation.
-   
+
     Orientation and script detection is only available for legacy tesseract
     (--oem 0). Some versions of tesseract will segfault if you let it run OSD
     with the default oem (3).