answer42 0.3.3__tar.gz → 0.3.4__tar.gz

{answer42-0.3.3 → answer42-0.3.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: answer42
-Version: 0.3.3
+Version: 0.3.4
 Summary: Answer42 — The Answer to Life, Universe, and 1C — UI Driver. MCP-powered 1C:Enterprise UI automation: click, fill, navigate, test, and introspect managed forms through the test-client API
 Author: Marvin (AI Assistant), 42Clouds, and contributors
 Author-email: "Kosolapov Stanislav (proDOOMman)" <prodoomman@gmail.com>

{answer42-0.3.3 → answer42-0.3.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "answer42"
-version = "0.3.3"
+version = "0.3.4"
 description = "Answer42 — The Answer to Life, Universe, and 1C — UI Driver. MCP-powered 1C:Enterprise UI automation: click, fill, navigate, test, and introspect managed forms through the test-client API"
 readme = "README.md"
 requires-python = ">=3.11"

{answer42-0.3.3 → answer42-0.3.4}/src/mcp_1c/rag/parsers.py

@@ -21,6 +21,8 @@ from .model import (
 )
 
 KIND_DIRS = {
+    "Constants": "Константа",
+    "CommonForms": "ОбщаяФорма",
     "Catalogs": "Справочник",
     "Documents": "Документ",
     "DataProcessors": "Обработка",
@@ -33,6 +35,8 @@ KIND_DIRS = {
     "CalculationRegisters": "РегистрРасчета",
 }
 
+FORM_FILE_NAMES = {"Form.form", "Form.xml"}
+
 
 def _strip_ns(tag: str) -> str:
     return tag.rsplit("}", 1)[-1]
@@ -199,7 +203,10 @@ def parse_edt_source(source: SourceInfo) -> ParsedSource:
             parsed.objects.append(MetadataObject(full, ru_kind, obj_name, synonym, str(mdo), source.format, hierarchical, [o for o in owners if o]))
             if root is not None:
                 _extract_metadata_children(parsed, root, full, mdo)
-            _extract_forms(parsed, mdo.parent, full)
+            if ru_kind == "ОбщаяФорма":
+                _extract_common_form(parsed, mdo.parent, full)
+            else:
+                _extract_forms(parsed, mdo.parent, full)
             _extract_data_composition_schemas(parsed, mdo.parent, full)
             _extract_help_pages(parsed, mdo.parent, full)
     return parsed
@@ -231,12 +238,26 @@ def parse_designer_xml_source(source: SourceInfo) -> ParsedSource:
             if root is not None:
                 _extract_metadata_children(parsed, root, full, xml)
             object_dir = xml.parent if xml.parent != base else base / xml.stem
-            _extract_forms(parsed, object_dir, full)
+            if ru_kind == "ОбщаяФорма":
+                _extract_common_form(parsed, object_dir, full)
+            else:
+                _extract_forms(parsed, object_dir, full)
             _extract_data_composition_schemas(parsed, object_dir, full)
             _extract_help_pages(parsed, object_dir, full)
     return parsed
 
 
+def _extract_common_form(parsed: ParsedSource, object_dir: Path, full: str) -> None:
+    form_xml = _find_form_file(object_dir)
+    parsed.forms.append(Form(full, "Форма", "CommonForm", str(form_xml) if form_xml.exists() else str(object_dir)))
+    if not form_xml.exists():
+        return
+    root = _xml_root(form_xml)
+    if root is not None:
+        _extract_form_elements(parsed, root, full, "Форма", form_xml)
+        _extract_embedded_data_composition_fields(parsed, root, full, "Форма", form_xml)
+
+
 def _extract_metadata_children(parsed: ParsedSource, root: ET.Element, full: str, source_path: Path) -> None:
     # Heuristic parser covering both EDT and designer XML structures.
     # Keep object attributes and table-part columns separate: XML dumps often
@@ -273,9 +294,7 @@ def _extract_forms(parsed: ParsedSource, object_dir: Path, full: str) -> None:
         return
     for form_dir in sorted([p for p in forms_dir.iterdir() if p.is_dir()]):
         form_name = form_dir.name
-        form_xml = form_dir / "Ext" / "Form.xml"
-        if not form_xml.exists():
-            form_xml = form_dir / "Form.xml"
+        form_xml = _find_form_file(form_dir)
         parsed.forms.append(Form(full, form_name, None, str(form_xml) if form_xml.exists() else str(form_dir)))
         if form_xml.exists():
             root = _xml_root(form_xml)
@@ -284,23 +303,34 @@ def _extract_forms(parsed: ParsedSource, object_dir: Path, full: str) -> None:
                 _extract_embedded_data_composition_fields(parsed, root, full, form_name, form_xml)
 
 
+def _find_form_file(form_dir: Path) -> Path:
+    for candidate_dir in (form_dir / "Ext", form_dir):
+        for name in FORM_FILE_NAMES:
+            form_file = candidate_dir / name
+            if form_file.exists():
+                return form_file
+    return form_dir / "Ext" / "Form.xml"
+
+
 def _extract_form_elements(parsed: ParsedSource, root: ET.Element, full: str, form_name: str, source_path: Path) -> None:
     for elem in root.iter():
         tag = _strip_ns(elem.tag)
         low = tag.lower()
-        if low not in {"element", "item", "items", "button", "command", "field", "table", "group"}:
+        if low not in {"element", "item", "items", "button", "command", "field", "table", "group", "extendedtooltip", "contextmenu"}:
             continue
         name = elem.attrib.get("name") or elem.attrib.get("Name") or _first_text(elem, ["name", "Name"])
         if not name:
             continue
+        title = elem.attrib.get("title") or elem.attrib.get("Title") or _first_text(elem, ["title", "Title", "caption", "Caption"])
+        data_path = _form_data_path(elem)
         parsed.form_elements.append(
             FormElement(
                 full,
                 form_name,
                 name,
                 tag,
-                elem.attrib.get("title") or elem.attrib.get("Title") or _first_text(elem, ["title", "Title", "caption", "Caption"]),
-                elem.attrib.get("dataPath") or elem.attrib.get("path") or _first_text(elem, ["dataPath", "path", "DataPath"]),
+                title,
+                data_path,
                 elem.attrib.get("commandName") or _first_text(elem, ["commandName", "CommandName"]),
                 None,
                 str(source_path),
@@ -308,6 +338,20 @@ def _extract_form_elements(parsed: ParsedSource, root: ET.Element, full: str, fo
         )
 
 
+def _form_data_path(elem: ET.Element) -> str | None:
+    direct = elem.attrib.get("dataPath") or elem.attrib.get("path") or _direct_text(elem, ["dataPath", "path", "DataPath"])
+    if direct:
+        return direct
+    for child in list(elem):
+        if _strip_ns(child.tag).lower() == "datapath":
+            segments = [_text(segment) for segment in child.iter() if _strip_ns(segment.tag).lower() in {"segments", "segment"}]
+            values = [value for value in segments if value]
+            if values:
+                return ".".join(values)
+            return _text(child)
+    return None
+
+
 def _extract_embedded_data_composition_fields(parsed: ParsedSource, root: ET.Element, full: str, form_name: str, source_path: Path) -> None:
     dataset_name: str | None = None
     for elem in root.iter():

{answer42-0.3.3 → answer42-0.3.4}/src/mcp_1c/rag/service.py

@@ -266,6 +266,12 @@ class RagService:
                         continue
                     con.execute("INSERT INTO attributes(object_id,name,type_name,synonym,required,source_path,role,raw_json) VALUES(?,?,?,?,?,?,?,?)", (oid, attr.name, attr.type_name, attr.synonym, attr.required, attr.source_path, attr.role, json_dumps(attr.raw)))
                     counts["attributes"] += 1
+                    text_parts = [attr.object_full_name, attr.role, attr.name, search_text_variants(attr.name, attr.synonym), attr.synonym or "", attr.type_name or ""]
+                    chunk_text = " ".join(part for part in text_parts if part)
+                    cid = _insert_chunk(con, snapshot_id, oid, "attribute", chunk_text, attr.source_path, {"source": src["name"], "role": attr.role})
+                    con.execute("INSERT INTO search_fts(text, object_full_name, kind, source_path, chunk_id) VALUES(?,?,?,?,?)", (chunk_text, attr.object_full_name, "Реквизит", attr.source_path, cid))
+                    _insert_embedding(con, cid, chunk_text)
+                    counts["chunks"] += 1
                 for tp in parsed.table_parts:
                     oid = object_ids.get(tp.object_full_name)
                     if not oid:
@@ -284,12 +290,25 @@ class RagService:
                     cur = con.execute("INSERT INTO forms(object_id,name,kind,source_path,raw_json) VALUES(?,?,?,?,?)", (oid, form.name, form.kind, form.source_path, json_dumps(form.raw)))
                     form_ids[(form.object_full_name, form.name)] = cur.lastrowid
                     counts["forms"] += 1
+                    text_parts = [form.object_full_name, "форма", form.name, search_text_variants(form.name), form.kind or ""]
+                    chunk_text = " ".join(part for part in text_parts if part)
+                    cid = _insert_chunk(con, snapshot_id, oid, "form", chunk_text, form.source_path, {"source": src["name"], "form": form.name})
+                    con.execute("INSERT INTO search_fts(text, object_full_name, kind, source_path, chunk_id) VALUES(?,?,?,?,?)", (chunk_text, form.object_full_name, "Форма", form.source_path, cid))
+                    _insert_embedding(con, cid, chunk_text)
+                    counts["chunks"] += 1
                 for elem in parsed.form_elements:
                     fid = form_ids.get((elem.object_full_name, elem.form_name))
                     if not fid:
                         continue
                     con.execute("INSERT INTO form_elements(form_id,name,element_type,title,data_path,command_name,parent_name,source_path,raw_json) VALUES(?,?,?,?,?,?,?,?,?)", (fid, elem.name, elem.element_type, elem.title, elem.data_path, elem.command_name, elem.parent_name, elem.source_path, json_dumps(elem.raw)))
                     counts["form_elements"] += 1
+                    oid = object_ids.get(elem.object_full_name)
+                    text_parts = [elem.object_full_name, "элемент формы", elem.form_name, elem.name, search_text_variants(elem.name, elem.title, elem.data_path), elem.title or "", elem.data_path or "", elem.command_name or "", elem.element_type or ""]
+                    chunk_text = " ".join(part for part in text_parts if part)
+                    cid = _insert_chunk(con, snapshot_id, oid, "form_element", chunk_text, elem.source_path, {"source": src["name"], "form": elem.form_name, "element": elem.name})
+                    con.execute("INSERT INTO search_fts(text, object_full_name, kind, source_path, chunk_id) VALUES(?,?,?,?,?)", (chunk_text, elem.object_full_name, "ЭлементФормы", elem.source_path, cid))
+                    _insert_embedding(con, cid, chunk_text)
+                    counts["chunks"] += 1
                 for dcs_field in parsed.data_composition_fields:
                     oid = object_ids.get(dcs_field.object_full_name)
                     if not oid: