switch to code formula model v1.0.1 and new test pdf

Signed-off-by: Matteo-Omenetti <[email protected]>
DS4SD · Feb 4, 2025 · 68d1713 · 68d1713
1 parent 89844a5
commit 68d1713
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 5 deletions.
diff --git a/docling/models/code_formula_model.py b/docling/models/code_formula_model.py
@@ -62,7 +62,7 @@ class CodeFormulaModel(BaseItemAndImageEnrichmentModel):
     """
 
     images_scale = 1.66  # = 120 dpi, aligned with training data resolution
-    expansion_factor = 0.05
+    expansion_factor = 0.03
 
     def __init__(
         self,

diff --git a/tests/data/code_and_formula.pdf b/tests/data/code_and_formula.pdf
diff --git a/tests/test_code_formula.py b/tests/test_code_formula.py
@@ -48,11 +48,11 @@ def test_code_and_formula_conversion():
     code_blocks = [el for el in results if isinstance(el, CodeItem)]
     assert len(code_blocks) == 1
 
-    gt = 'int main() {\n    printf("Hello, World!");\n    return 0;\n}'
+    gt = 'function add(a, b) {\n    return a + b;\n}\nconsole.log(add(3, 5));'
 
-    predicted = code_blocks[0].text.strip()
+    predicted = code_blocks[0].text.strip() 
     assert predicted == gt, f"mismatch in text {predicted=}, {gt=}"
-    assert code_blocks[0].code_language == CodeLanguageLabel.C_PLUS_PLUS
+    assert code_blocks[0].code_language == CodeLanguageLabel.JAVASCRIPT
 
     formula_blocks = [
         el
@@ -61,6 +61,6 @@ def test_code_and_formula_conversion():
     ]
     assert len(formula_blocks) == 1
 
-    gt = "a ^ { 2 } + 8 = 1 2"
+    gt = "x ^ { 2 } + 8 = 1 2"
     predicted = formula_blocks[0].text
     assert predicted == gt, f"mismatch in text {predicted=}, {gt=}"