updated how we are doing manifests so they are more clean and easier to check file formats

fccs_export.py

@@ -168,20 +168,62 @@ def open_send_to_file(main, app, ctrl, timeouts):
     return dlg
 
 
-def read_manifest(dlg, log):
-    """Read the list of selected documents from the right-side ListView."""
+def read_manifest(dlg, drawer_id, log):
+    """Read the selected documents from the ListView into clean rows.
+
+    The ListView (`List1`) has columns Drawer ID, Page Title (the document
+    name) and Application. `texts()` returns a flat, row-major dump led by the
+    control's own name ('List1'), which is why the raw capture looked messy.
+
+    This parses it into one row per document and drops the redundant Drawer ID
+    column (always equal to drawer_id) and any empty columns, leaving the
+    document name (and Application) per line. Returns a list of rows (each a
+    list of cell strings), [] if empty, or None if it couldn't be read.
+    """
     try:
         lv = dlg.child_window(title="List1", class_name="SysListView32")
-        count = lv.item_count()
-        if count == 0:
+        n_rows = lv.item_count()
+    except Exception as e:
+        log(f"  WARNING: could not read manifest list: {e}")
+        return None
+
+    if n_rows == 0:
         log("  WARNING: ListView is empty after Select")
         return []
-        items = lv.texts()
-        log(f"  Manifest captured: {count} documents selected")
-        return items
+
+    try:
+        n_cols = lv.column_count()
+    except Exception:
+        n_cols = 0
+    if not n_cols or n_cols < 1:
+        n_cols = 3  # observed columns: Drawer ID, Page Title, Application
+
+    # Read every cell. Primary: get_item(row, col). Fallback: reshape texts().
+    try:
+        grid = [[lv.get_item(r, c).text() for c in range(n_cols)]
+                for r in range(n_rows)]
     except Exception as e:
-        log(f"  WARNING: could not read manifest from ListView: {e}")
-        return None
+        log(f"  NOTE: get_item failed ({e}); falling back to texts().")
+        raw = lv.texts()
+        if raw and raw[0] == lv.window_text():
+            raw = raw[1:]  # drop the control's own name ('List1')
+        grid = [raw[i:i + n_cols] for i in range(0, len(raw), n_cols)]
+        grid = [row for row in grid if len(row) == n_cols]
+
+    # Keep only columns that carry real, non-redundant info.
+    keep_cols = []
+    for c in range(n_cols):
+        vals = [(row[c].strip() if c < len(row) else "") for row in grid]
+        if all(v == drawer_id for v in vals):
+            continue  # redundant Drawer ID column
+        if all(v == "" for v in vals):
+            continue  # empty column
+        keep_cols.append(c)
+
+    rows = [[(row[c] if c < len(row) else "") for c in keep_cols]
+            for row in grid]
+    log(f"  Manifest captured: {len(rows)} documents selected")
+    return rows
 
 
 def save_manifest(drawer_id, items, manifest_dir, log):
@@ -197,8 +239,8 @@ def save_manifest(drawer_id, items, manifest_dir, log):
     log(f"  Manifest saved: {path}")
 
 
-def perform_export(dlg, ctrl, timeouts, log):
-    """Click Select -> then OK to start the export. Returns manifest list."""
+def perform_export(dlg, drawer_id, ctrl, timeouts, log):
+    """Click Select -> then OK to start the export. Returns manifest rows."""
     select_btn = dlg.child_window(title=ctrl["select_btn_title"], class_name="Button")
     select_btn.wait("visible ready enabled", timeout=timeouts["dialog"])
     select_btn.click()
@@ -206,7 +248,7 @@ def perform_export(dlg, ctrl, timeouts, log):
     ok_btn = dlg.child_window(title=ctrl["ok_btn_title"], class_name="Button")
     ok_btn.wait("visible ready enabled", timeout=timeouts["select"])
 
-    manifest = read_manifest(dlg, log)
+    manifest = read_manifest(dlg, drawer_id, log)
 
     ok_btn.click()
     return manifest
@@ -319,7 +361,7 @@ def export_drawer(app, main, search_box, drawer_id, ctrl, timeouts,
 
     # 3 & 4. Select -> and OK (captures manifest before clicking OK)
     try:
-        manifest = perform_export(dlg, ctrl, timeouts, log)
+        manifest = perform_export(dlg, drawer_id, ctrl, timeouts, log)
         if manifest is not None:
             save_manifest(drawer_id, manifest, manifest_dir, log)
     except Exception as e:

fccs_verify.py

@@ -13,14 +13,23 @@ from fccs_config import parse_args, load_config, make_logger
 
 
 def load_manifest(path):
-    """Load a manifest file and return the list of document rows."""
-    items = []
+    """Load a manifest file and return one row per document.
+
+    Handles both formats:
+      - New: one document per line (tab-separated columns).
+      - Old: a raw ListView dump led by 'List1', then row-major cells with
+        3 columns per document (Drawer ID, Page Title, Application). We
+        reshape it so the document count is correct.
+    """
     with open(path, "r", encoding="utf-8") as f:
-        for line in f:
-            line = line.rstrip("\n")
-            if line:
-                items.append(line.split("\t"))
-    return items
+        lines = [line.rstrip("\n") for line in f if line.strip()]
+
+    if lines and lines[0].strip() == "List1":
+        cells = lines[1:]
+        rows = [cells[i:i + 3] for i in range(0, len(cells), 3)]
+        return [r for r in rows if len(r) == 3]
+
+    return [line.split("\t") for line in lines]
 
 
 def main():