diff options
| author | GeorgeFI | 2023-09-25 21:12:26 +0200 |
|---|---|---|
| committer | GeorgeFI | 2023-09-25 21:12:26 +0200 |
| commit | 498308a737943b87ecbf38a05404d4caa22fa726 (patch) | |
| tree | 3f1fcf4264f493da46dcb24a8ea21024c502ccbd /src | |
| parent | 0eaf610cfe6a0a36cfad2fa71b42d1f921b6e49b (diff) | |
| download | sec-certs-498308a737943b87ecbf38a05404d4caa22fa726.tar.gz sec-certs-498308a737943b87ecbf38a05404d4caa22fa726.tar.zst sec-certs-498308a737943b87ecbf38a05404d4caa22fa726.zip | |
fix: OCR with Pytesseract fixed now
Diffstat (limited to 'src')
| -rw-r--r-- | src/sec_certs/utils/pdf.py | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/src/sec_certs/utils/pdf.py b/src/sec_certs/utils/pdf.py index b4ec1e94..7736b5d0 100644 --- a/src/sec_certs/utils/pdf.py +++ b/src/sec_certs/utils/pdf.py @@ -12,6 +12,7 @@ from typing import Any import pdftotext import pikepdf import pytesseract +from PIL import Image from sec_certs import constants from sec_certs.constants import ( @@ -55,7 +56,7 @@ def ocr_pdf_file(pdf_path: Path) -> str: for ppm_path in map(Path, glob.glob(str(tmppath / "image*.ppm"))): base = ppm_path.with_suffix("") - content = pytesseract.image_to_string(ppm_path, lang="eng+deu+fra") + content = pytesseract.image_to_string(Image.open(ppm_path), lang="eng+deu+fra") if content: with Path(base.with_suffix(".txt")).open("w") as file: |