Add page-marker grouping and party-aware naming

Consecutive captures whose printed PAGE X OF Y markers advance under the same total are grouped into one multi-page PDF, with duplicate pages keeping the better read. Naming gains doc-title/party components and a scan-date fallback so nothing lands in UNSORTED. Vision LLM default switches to minicpm-v (faster, fewer false blanks on dense mono pages).
2026-07-26 15:37:28 -04:00
parent b11f46b00f
commit 12e5d1aeb1
10 changed files with 986 additions and 135 deletions
@@ -102,9 +102,10 @@ class OcrConfig:
 class LlmConfig:
    # Use a local Ollama vision model to identify/name documents and flag
    # quality issues. Falls back to Tesseract heuristics when Ollama is
-    # down or the model errors. ~20-40s per document vs <2s for Tesseract.
+    # down or the model errors. minicpm-v is ~8-12s/doc and much less prone
+    # to false "blank" labels than qwen2.5vl on dense statement pages.
    enabled: bool = True
-    model: str = "qwen2.5vl"
+    model: str = "minicpm-v"
    base_url: str = "http://localhost:11434"
    timeout_s: float = 180.0