Skip to content

Commit 208d9dd

Browse files
authored
Setting DOCLING_ARTIFACTS_PATH to hardcode RapidOCR downloads (#1147)
1 parent 56a6635 commit 208d9dd

File tree

2 files changed

+23
-20
lines changed

2 files changed

+23
-20
lines changed

.github/workflows/tests.yml

Lines changed: 5 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -110,11 +110,14 @@ jobs:
110110
~/.cache/docling
111111
key: ${{ runner.os }}-docling-${{ hashFiles('uv.lock') }}
112112
restore-keys: ${{ runner.os }}-docling-
113-
- name: Pre-download Docling's RapidOCR models # Avoid CI race conditions in filesystem on model download
113+
- name: Pre-download Docling models # Avoid CI race conditions in filesystem on model download
114114
if: steps.cache-models.outputs.cache-hit != 'true'
115-
run: uv run docling-tools models download rapidocr
115+
# RapidOCR is used for PDF pipeline's OCR, layout for layout analysis,
116+
# tableformer for table structure
117+
run: uv run docling-tools models download rapidocr layout tableformer
116118
- run: uv run pytest -n auto packages
117119
env:
120+
DOCLING_ARTIFACTS_PATH: ~/.cache/docling/models # Work around https://github.com/docling-project/docling/issues/2500
118121
OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
119122
ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
120123
GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}

uv.lock

Lines changed: 18 additions & 18 deletions
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

0 commit comments

Comments
 (0)