Added support for using Apache Tika as a document loader.

Added persistent configuration options to configure use and location of Tika service.

Updated backend.apps.rag.main:get_loader() to make use of Tika document loader.

.dockerignore

@@ -10,7 +10,7 @@ node_modules
 vite.config.js.timestamp-*
 vite.config.ts.timestamp-*
 __pycache__
-.env
+.idea
 _old
 uploads
 .ipynb_checkpoints

backend/apps/rag/main.py

@@ -93,6 +93,8 @@ from config import (
     SRC_LOG_LEVELS,
     UPLOAD_DIR,
     DOCS_DIR,
+    DOCUMENT_USE_TIKA,
+    TIKA_SERVER_URL,
     RAG_TOP_K,
     RAG_RELEVANCE_THRESHOLD,
     RAG_EMBEDDING_ENGINE,
@@ -985,6 +987,41 @@ def store_docs_in_vector_db(docs, collection_name, overwrite: bool = False) -> b
         return False
 
 
+class TikaLoader:
+    def __init__(self, file_path, mime_type=None):
+        self.file_path = file_path
+        self.mime_type = mime_type
+
+    def load(self) -> List[Document]:
+        with (open(self.file_path, "rb") as f):
+            data = f.read()
+
+        if self.mime_type is not None:
+            headers = {"Content-Type": self.mime_type}
+        else:
+            headers = {}
+
+        endpoint = str(TIKA_SERVER_URL)
+        if not endpoint.endswith("/"):
+            endpoint += "/"
+        endpoint += "tika/text"
+
+        r = requests.put(endpoint, data=data, headers=headers)
+
+        if r.ok:
+            raw_metadata = r.json()
+            text = raw_metadata.get("X-TIKA:content", "<No text content found>")
+
+            if "Content-Type" in raw_metadata:
+                headers["Content-Type"] = raw_metadata["Content-Type"]
+
+            log.info("Tika extracted text: %s", text)
+
+            return [Document(page_content=text, metadata=headers)]
+        else:
+            raise Exception(f"Error calling Tika: {r.reason}")
+
+
 def get_loader(filename: str, file_content_type: str, file_path: str):
     file_ext = filename.split(".")[-1].lower()
     known_type = True
@@ -1035,6 +1072,16 @@ def get_loader(filename: str, file_content_type: str, file_path: str):
         "msg",
     ]
 
+    log.warning("Use tika: %s, server URL: %s", DOCUMENT_USE_TIKA, TIKA_SERVER_URL)
+
+    if DOCUMENT_USE_TIKA and TIKA_SERVER_URL:
+        if file_ext in known_source_ext or (
+                file_content_type and file_content_type.find("text/") >= 0
+        ):
+            loader = TextLoader(file_path, autodetect_encoding=True)
+        else:
+            loader = TikaLoader(file_path, file_content_type)
+    else:
         if file_ext == "pdf":
             loader = PyPDFLoader(
                 file_path, extract_images=app.state.config.PDF_EXTRACT_IMAGES

backend/config.py

@@ -878,6 +878,22 @@ WEBUI_SESSION_COOKIE_SECURE = os.environ.get(
 if WEBUI_AUTH and WEBUI_SECRET_KEY == "":
     raise ValueError(ERROR_MESSAGES.ENV_VAR_NOT_FOUND)
 
+####################################
+# RAG document text extraction
+####################################
+
+DOCUMENT_USE_TIKA = PersistentConfig(
+    "DOCUMENT_USE_TIKA",
+    "rag.document_use_tika",
+    os.environ.get("DOCUMENT_USE_TIKA", "false").lower() == "true"
+)
+
+TIKA_SERVER_URL = PersistentConfig(
+    "TIKA_SERVER_URL",
+    "rag.tika_server_url",
+    os.getenv("TIKA_SERVER_URL", "http://tika:9998"),  # Default for sidecar deployment
+)
+
 ####################################
 # RAG
 ####################################