debase 0.5.0__tar.gz → 0.5.1__tar.gz

{debase-0.5.0/src/debase.egg-info → debase-0.5.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: debase
-Version: 0.5.0
+Version: 0.5.1
 Summary: Enzyme lineage analysis and sequence extraction package
 Home-page: https://github.com/YuemingLong/DEBase
 Author: DEBase Team

{debase-0.5.0 → debase-0.5.1}/src/debase/_version.py

@@ -1,3 +1,3 @@
 """Version information."""
 
-__version__ = "0.5.0"
+__version__ = "0.5.1"

{debase-0.5.0 → debase-0.5.1}/src/debase/reaction_info_extractor.py

@@ -1195,7 +1195,8 @@ class ReactionExtractor:
                 # Create a flexible pattern that handles various spacing and formatting
                 # This pattern looks for "Figure" (case insensitive) followed by optional spaces
                 # then the figure number, then any of: period, colon, space+capital letter, or end of line
-                flexible_pattern = rf"(?i)figure\s*{re.escape(main_figure_num)}(?:\.|:|(?=\s+[A-Z])|\s*$)"
+                # Also match at the beginning of a line to catch captions
+                flexible_pattern = rf"(?i)(?:^|\n)\s*figure\s*{re.escape(main_figure_num)}(?:\.|:|(?=\s+[A-Z])|\s*$)"
                 
                 LOGGER.debug("Looking for figure caption '%s' with flexible pattern: %s", 
                            main_figure_num, flexible_pattern)
@@ -1231,11 +1232,17 @@ class ReactionExtractor:
                         continue
                         
                     # Check if the remaining text looks like a caption (contains descriptive words)
+                    # Expanded list of caption keywords to be more inclusive
                     first_words = remaining_text[:50].lower()
-                    if not any(word in first_words for word in ['detailed', 'representative', 'shows', 'comparison', 
-                                                                 'illustrates', 'demonstrates', 'results', 'data',
-                                                                 'chromatogram', 'spectra', 'analysis', 'site-directed',
-                                                                 'mutagenesis', 'mutants']):
+                    caption_keywords = ['detailed', 'representative', 'shows', 'comparison', 
+                                      'illustrates', 'demonstrates', 'results', 'data',
+                                      'chromatogram', 'spectra', 'analysis', 'site-directed',
+                                      'mutagenesis', 'mutants', 'evolution', 'directed',
+                                      'screening', 'reaction', 'variant', 'enzyme', 'protein',
+                                      'activity', 'performance', 'yield', 'selectivity',
+                                      'characterization', 'optimization', 'development',
+                                      'structure', 'domain', 'crystal', 'model']
+                    if not any(word in first_words for word in caption_keywords):
                         LOGGER.debug("Skipping: doesn't look like caption text: %s", first_words)
                         continue
                     
@@ -1322,6 +1329,39 @@ class ReactionExtractor:
                     self._figure_cache.put(cache_key, result)
                     return result
         
+        # Fallback: If no caption found, try to find any page that mentions this figure
+        LOGGER.info("No figure caption found for '%s', trying fallback search", ref)
+        
+        for doc_idx, doc in enumerate(docs):
+            doc_name = "MS" if doc_idx == 0 else "SI"
+            for page_number in range(doc.page_count):
+                page = doc.load_page(page_number)
+                page_text = page.get_text()
+                
+                # Look for any mention of the figure reference
+                if re.search(rf'\b{re.escape(ref)}\b', page_text, re.IGNORECASE):
+                    LOGGER.info("Found '%s' mentioned on page %d of %s document (fallback)", 
+                               ref, page_number + 1, doc_name)
+                    
+                    # Extract the entire page as the figure might be on this page
+                    mat = fitz.Matrix(5.0, 5.0)  # 5x zoom for better quality
+                    pix = page.get_pixmap(matrix=mat)
+                    pix = self._ensure_rgb_pixmap(pix)
+                    img_bytes = pix.tobytes("png")
+                    
+                    # Save PNG to debug directory if available
+                    if self.debug_dir:
+                        timestamp = int(time.time())
+                        png_file = self.debug_dir / f"fallback_{ref.replace(' ', '_')}_{timestamp}.png"
+                        with open(png_file, 'wb') as f:
+                            f.write(img_bytes)
+                        LOGGER.info("Saved fallback page image to: %s", png_file)
+                    
+                    result = b64encode(img_bytes).decode()
+                    # Cache the result
+                    self._figure_cache.put(cache_key, result)
+                    return result
+        
         LOGGER.warning("_extract_page_png returning None for '%s' - figure not found in any document", ref)
         return None
     

{debase-0.5.0 → debase-0.5.1/src/debase.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: debase
-Version: 0.5.0
+Version: 0.5.1
 Summary: Enzyme lineage analysis and sequence extraction package
 Home-page: https://github.com/YuemingLong/DEBase
 Author: DEBase Team