Fix: Tika server upgrades. (#12073)

### What problem does this PR solve?

#12037

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)

api/apps/chunk_app.py

@@ -76,6 +76,7 @@ async def list_chunk():
                 "image_id": sres.field[id].get("img_id", ""),
                 "available_int": int(sres.field[id].get("available_int", 1)),
                 "positions": sres.field[id].get("position_int", []),
+                "doc_type_kwd": sres.field[id].get("doc_type_kwd")
             }
             assert isinstance(d["positions"], list)
             assert len(d["positions"]) == 0 or (isinstance(d["positions"][0], list) and len(d["positions"][0]) == 5)
@@ -176,10 +177,9 @@ async def set():
             settings.docStoreConn.update({"id": req["chunk_id"]}, _d, search.index_name(tenant_id), doc.kb_id)
 
             # update image
-            image_id = req.get("img_id")
-            bkt, name = image_id.split("-")
             image_base64 = req.get("image_base64", None)
             if image_base64:
+                bkt, name = req.get("img_id", "-").split("-")
                 image_binary = base64.b64decode(image_base64)
                 settings.STORAGE_IMPL.put(bkt, name, image_binary)
             return get_json_result(data=True)

api/apps/kb_app.py

@@ -150,6 +150,21 @@ async def update():
         return server_error_response(e)
 
 
+@manager.route('/update_metadata_setting', methods=['post'])  # noqa: F821
+@login_required
+@validate_request("kb_id", "metadata")
+async def update_metadata_setting():
+    req = await get_request_json()
+    e, kb = KnowledgebaseService.get_by_id(req["kb_id"])
+    if not e:
+        return get_data_error_result(
+            message="Database error (Knowledgebase rename)!")
+    kb = kb.to_dict()
+    kb["parser_config"]["metadata"] = req["metadata"]
+    KnowledgebaseService.update_by_id(kb["id"], kb)
+    return get_json_result(data=kb)
+
+
 @manager.route('/detail', methods=['GET'])  # noqa: F821
 @login_required
 def detail():

api/db/services/knowledgebase_service.py

@@ -411,6 +411,8 @@ class KnowledgebaseService(CommonService):
         ok, _t = TenantService.get_by_id(tenant_id)
         if not ok:
             return False, get_data_error_result(message="Tenant not found.")
+        if kwargs.get("parser_config") and isinstance(kwargs["parser_config"], dict) and not kwargs["parser_config"].get("llm_id"):
+            kwargs["parser_config"]["llm_id"] = _t.llm_id
 
         # Build payload
         kb_id = get_uuid()

conf/infinity_mapping.json

@@ -31,6 +31,7 @@
 	"entity_type_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},
 	"source_id": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},
 	"n_hop_with_weight": {"type": "varchar", "default": ""},
+	"mom_with_weight": {"type": "varchar", "default": ""},
 	"removed_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},
 	"doc_type_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},
 	"toc_kwd": {"type": "varchar", "default": "", "analyzer": "whitespace-#"},

download_deps.py

@@ -23,8 +23,8 @@ def get_urls(use_china_mirrors=False) -> list[Union[str, list[str]]]:
         return [
             "http://mirrors.tuna.tsinghua.edu.cn/ubuntu/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2_amd64.deb",
             "http://mirrors.tuna.tsinghua.edu.cn/ubuntu-ports/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2_arm64.deb",
-            "https://repo.huaweicloud.com/repository/maven/org/apache/tika/tika-server-standard/3.0.0/tika-server-standard-3.0.0.jar",
+            "https://repo.huaweicloud.com/repository/maven/org/apache/tika/tika-server-standard/3.2.3/tika-server-standard-3.2.3.jar",
-            "https://repo.huaweicloud.com/repository/maven/org/apache/tika/tika-server-standard/3.0.0/tika-server-standard-3.0.0.jar.md5",
+            "https://repo.huaweicloud.com/repository/maven/org/apache/tika/tika-server-standard/3.2.3/tika-server-standard-3.2.3.jar.md5",
             "https://openaipublic.blob.core.windows.net/encodings/cl100k_base.tiktoken",
             ["https://registry.npmmirror.com/-/binary/chrome-for-testing/121.0.6167.85/linux64/chrome-linux64.zip", "chrome-linux64-121-0-6167-85"],
             ["https://registry.npmmirror.com/-/binary/chrome-for-testing/121.0.6167.85/linux64/chromedriver-linux64.zip", "chromedriver-linux64-121-0-6167-85"],
@@ -34,8 +34,8 @@ def get_urls(use_china_mirrors=False) -> list[Union[str, list[str]]]:
         return [
             "http://archive.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2_amd64.deb",
             "http://ports.ubuntu.com/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2_arm64.deb",
-            "https://repo1.maven.org/maven2/org/apache/tika/tika-server-standard/3.0.0/tika-server-standard-3.0.0.jar",
+            "https://repo1.maven.org/maven2/org/apache/tika/tika-server-standard/3.2.3/tika-server-standard-3.2.3.jar",
-            "https://repo1.maven.org/maven2/org/apache/tika/tika-server-standard/3.0.0/tika-server-standard-3.0.0.jar.md5",
+            "https://repo1.maven.org/maven2/org/apache/tika/tika-server-standard/3.2.3/tika-server-standard-3.2.3.jar.md5",
             "https://openaipublic.blob.core.windows.net/encodings/cl100k_base.tiktoken",
             ["https://storage.googleapis.com/chrome-for-testing-public/121.0.6167.85/linux64/chrome-linux64.zip", "chrome-linux64-121-0-6167-85"],
             ["https://storage.googleapis.com/chrome-for-testing-public/121.0.6167.85/linux64/chromedriver-linux64.zip", "chromedriver-linux64-121-0-6167-85"],

rag/app/book.py

@@ -91,7 +91,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             filename, binary=binary, from_page=from_page, to_page=to_page)
         remove_contents_table(sections, eng=is_english(
             random_choices([t for t, _ in sections], k=200)))
-        tbls=vision_figure_parser_docx_wrapper(sections=sections,tbls=tbls,callback=callback,**kwargs)
+        tbls = vision_figure_parser_docx_wrapper(sections=sections,tbls=tbls,callback=callback,**kwargs)
         # tbls = [((None, lns), None) for lns in tbls]
         sections=[(item[0],item[1] if item[1] is not None else "") for item in sections if not isinstance(item[1], Image.Image)]
         callback(0.8, "Finish parsing.")
@@ -147,9 +147,16 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
 
     elif re.search(r"\.doc$", filename, re.IGNORECASE):
         callback(0.1, "Start to parse.")
-        with BytesIO(binary) as binary:
+        try:
-            binary = BytesIO(binary)
+            from tika import parser as tika_parser
-            doc_parsed = parser.from_buffer(binary)
+        except Exception as e:
+            callback(0.8, f"tika not available: {e}. Unsupported .doc parsing.")
+            logging.warning(f"tika not available: {e}. Unsupported .doc parsing for {filename}.")
+            return []
+
+        binary = BytesIO(binary)
+        doc_parsed = tika_parser.from_buffer(binary)
+        if doc_parsed.get('content', None) is not None:
             sections = doc_parsed['content'].split('\n')
             sections = [(line, "") for line in sections if line]
             remove_contents_table(sections, eng=is_english(

rag/app/manual.py

@@ -312,7 +312,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
             tk_cnt = num_tokens_from_string(txt)
             if sec_id > -1:
                 last_sid = sec_id
-        tbls=vision_figure_parser_pdf_wrapper(tbls=tbls,callback=callback,**kwargs)
+        tbls = vision_figure_parser_pdf_wrapper(tbls=tbls,callback=callback,**kwargs)
         res = tokenize_table(tbls, doc, eng)
         res.extend(tokenize_chunks(chunks, doc, eng, pdf_parser))
         table_ctx = max(0, int(parser_config.get("table_context_size", 0) or 0))
@@ -325,7 +325,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
         docx_parser = Docx()
         ti_list, tbls = docx_parser(filename, binary,
                                     from_page=0, to_page=10000, callback=callback)
-        tbls=vision_figure_parser_docx_wrapper(sections=ti_list,tbls=tbls,callback=callback,**kwargs)
+        tbls = vision_figure_parser_docx_wrapper(sections=ti_list,tbls=tbls,callback=callback,**kwargs)
         res = tokenize_table(tbls, doc, eng)
         for text, image in ti_list:
             d = copy.deepcopy(doc)

rag/app/one.py

@@ -76,7 +76,7 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
     if re.search(r"\.docx$", filename, re.IGNORECASE):
         callback(0.1, "Start to parse.")
         sections, tbls = naive.Docx()(filename, binary)
-        tbls=vision_figure_parser_docx_wrapper(sections=sections,tbls=tbls,callback=callback,**kwargs)
+        tbls = vision_figure_parser_docx_wrapper(sections=sections, tbls=tbls, callback=callback, **kwargs)
         sections = [s for s, _ in sections if s]
         for (_, html), _ in tbls:
             sections.append(html)
@@ -142,10 +142,18 @@ def chunk(filename, binary=None, from_page=0, to_page=100000,
 
     elif re.search(r"\.doc$", filename, re.IGNORECASE):
         callback(0.1, "Start to parse.")
+        try:
+            from tika import parser as tika_parser
+        except Exception as e:
+            callback(0.8, f"tika not available: {e}. Unsupported .doc parsing.")
+            logging.warning(f"tika not available: {e}. Unsupported .doc parsing for {filename}.")
+            return []
+
         binary = BytesIO(binary)
-        doc_parsed = parser.from_buffer(binary)
+        doc_parsed = tika_parser.from_buffer(binary)
-        sections = doc_parsed['content'].split('\n')
+        if doc_parsed.get('content', None) is not None:
-        sections = [s for s in sections if s]
+            sections = doc_parsed['content'].split('\n')
+            sections = [s for s in sections if s]
         callback(0.8, "Finish parsing.")
 
     else:

rag/flow/parser/parser.py

@@ -650,7 +650,7 @@ class Parser(ProcessBase):
             tmpf.flush()
             tmp_path = os.path.abspath(tmpf.name)
 
-            seq2txt_mdl = LLMBundle(self._canvas.get_tenant_id(), LLMType.SPEECH2TEXT)
+            seq2txt_mdl = LLMBundle(self._canvas.get_tenant_id(), LLMType.SPEECH2TEXT, llm_name=conf["llm_id"])
             txt = seq2txt_mdl.transcription(tmp_path)
 
             self.set_output("text", txt)

rag/svr/task_executor.py

@@ -374,7 +374,7 @@ async def build_chunks(task, progress_callback):
         chat_mdl = LLMBundle(task["tenant_id"], LLMType.CHAT, llm_name=task["llm_id"], lang=task["language"])
 
         async def gen_metadata_task(chat_mdl, d):
-            cached = get_llm_cache(chat_mdl.llm_name, d["content_with_weight"], "metadata")
+            cached = get_llm_cache(chat_mdl.llm_name, d["content_with_weight"], "metadata", {})
             if not cached:
                 async with chat_limiter:
                     cached = await gen_metadata(chat_mdl,
@@ -852,7 +852,7 @@ async def do_handle_task(task):
     task_tenant_id = task["tenant_id"]
     task_embedding_id = task["embd_id"]
     task_language = task["language"]
-    task_llm_id = task["llm_id"]
+    task_llm_id = task["parser_config"].get("llm_id") or task["llm_id"]
     task_dataset_id = task["kb_id"]
     task_doc_id = task["doc_id"]
     task_document_name = task["name"]