fix: Pandoc doesn't support DOC

open-contracting · Apr 5, 2024 · 9b2d8fa · 9b2d8fa
1 parent 185117f
commit 9b2d8fa
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/README.md b/README.md
@@ -12,7 +12,7 @@ Install [Popper](https://poppler.freedesktop.org) for its `pdftotext` command. F
 brew install poppler
 ```
 
-Install [Pandoc](https://pandoc.org) to convert DOC and DOCX to text. For example, on macOS:
+Install [Pandoc](https://pandoc.org) to convert DOCX to text. For example, on macOS:
 
 ```shell
 brew install pandoc
@@ -73,7 +73,7 @@ Download data, for example:
 
 ### General
 
-Transform PDF, DOC, DOCX, BMP, PNG and JPEG to text files:
+Transform DOCX, BMP, PNG, JPEG and PDF to text files:
 
 ```shell
 ./manage.py any2txt data/do

diff --git a/manage.py b/manage.py
@@ -687,14 +687,15 @@ def any2txt(indir, skip_existing):
                 continue
             click.echo(infile)
 
-            if suffix in (".doc", ".docx"):
+            if suffix == ".docx":
                 text = pypandoc.convert_file(infile, "plain")
             elif suffix == (".bmp", ".jpeg", ".png"):
                 text = pytesseract.image_to_string(infile)
             elif suffix == ".pdf":
                 text = "\n".join(pytesseract.image_to_string(i) for i in pdf2image.convert_from_path(infile, dpi=500))
             else:
-                raise NotImplementedError(suffix)
+                click.secho(f"Unsupported format ({suffix})", fg="yellow")
+                continue
 
             if text:
                 with outfile.open("w") as f: