Upload 6 files

backend/media_analysis/data_export.py

@@ -50,6 +50,22 @@ def normalize_records(records: list[dict]) -> tuple[list[dict], list[str]]:
                     item[key] = float(normalized) if "." in normalized else int(normalized)
                 elif cleaned:
                     warnings.append(f"Row {index}: {key} is not an unambiguous number: {value!r}")
+        # Separate class designations that were embedded in a standard range.
+        # This is a purely textual rearrangement: no thresholds or signs are inferred.
+        rng = item.get("standard_range")
+        if isinstance(rng, str):
+            m = re.fullmatch(r"\s*(.*?)\s+(Class\s+(?:I|II|III|IV|V))\s*", rng, re.I)
+            if m and m.group(1).strip():
+                extracted_range = m.group(1).strip()
+                extracted_class = "Class " + m.group(2).split()[-1].upper()
+                existing_class = item.get("class_label")
+                if not existing_class:
+                    item["class_label"] = extracted_class
+                    item["standard_range"] = extracted_range
+                elif re.sub(r"\s+", " ", str(existing_class)).strip().lower() == extracted_class.lower():
+                    item["standard_range"] = extracted_range
+                else:
+                    warnings.append(f"Row {index}: conflicting class_label and standard_range; preserved original")
         result.append(item)
     return result, warnings
 

backend/media_analysis/figure_analysis.py

@@ -158,12 +158,6 @@ def analyze_media_question(question: str, selected_papers: list[dict], *, enable
         records, normalization_warnings = normalize_records(records)
         result["structured_data"] = records
         result["normalization_warnings"] = normalization_warnings
-        from .record_checks import validate_extraction
-        records, validation_warnings, coverage = validate_extraction(records, items if evidence_mode != "vision" else [])
-        result["structured_data"] = records
-        result["validation_warnings"] = validation_warnings
-        result["extraction_coverage"] = coverage
-
         result["csv_attachment"] = csv_attachment(records, result.get("sources") or [])
         if warning:
             result["export_warning"] = warning

backend/media_analysis/vision_client.py

@@ -125,10 +125,7 @@ def analyze_images(question: str, evidence: list[dict], *, model: str | None = N
             "Use stable field names: parameter, unit, minimum, maximum, mean, std, "
             "standard_range, class_label, classification, percent_samples, references, "
             "estimated (boolean), extraction_method (printed_table|positioned_text|visual_estimate), "
-            "and source_quality. Keep source_quality brief (e.g., image_checked, ambiguous, unverified). "
-            "For class labels (Class I, II, III), populate class_label separately and "
-            "exclude the class name from standard_range. Never guess blank percentages as zero. "
-            "Preserve original printed parameter names, units, inequality "
+            "and source_quality. Preserve original printed parameter names, units, inequality "
             "signs, and numeric precision. Numeric values must be JSON numbers (no thousands "
             "separators); use *_printed fields if preservation of original formatting matters. "
             "For ambiguous OCR signs, compare against image; if still unclear use null and "