From 9494e998083220b9f575345ef402d8cab9cfe37d Mon Sep 17 00:00:00 2001
From: Gabriel Luiz Freitas Almeida <gabriel@langflow.org>
Date: Thu, 4 Sep 2025 18:32:29 -0300
Subject: [PATCH] Set tweaks to settings passed from the UI

---
 src/api/langflow_files.py | 58 +++++++++++++++++++++++++++++++++------
 1 file changed, 49 insertions(+), 9 deletions(-)

diff --git a/src/api/langflow_files.py b/src/api/langflow_files.py
index be0a0293..f51d2c17 100644
--- a/src/api/langflow_files.py
+++ b/src/api/langflow_files.py
@@ -2,21 +2,24 @@ from starlette.requests import Request
 from starlette.responses import JSONResponse
 
 from services.langflow_file_service import LangflowFileService
+from utils.logging_config import get_logger
+
+logger = get_logger(__name__)
 
 
 async def upload_user_file(
     request: Request, langflow_file_service: LangflowFileService, session_manager
 ):
     try:
-        print("[DEBUG] upload_user_file endpoint called")
+        logger.debug("upload_user_file endpoint called")
         form = await request.form()
         upload_file = form.get("file")
         if upload_file is None:
-            print("[ERROR] No file provided in upload request")
+            logger.error("No file provided in upload request")
             return JSONResponse({"error": "Missing file"}, status_code=400)
 
-        print(
-            f"[DEBUG] Processing file: {upload_file.filename}, size: {upload_file.size}"
+        logger.debug(
+            "Processing file", filename=upload_file.filename, size=upload_file.size
         )
 
         # starlette UploadFile provides file-like; httpx needs (filename, file, content_type)
@@ -28,16 +31,20 @@ async def upload_user_file(
         )
 
         jwt_token = getattr(request.state, "jwt_token", None)
-        print(f"[DEBUG] JWT token present: {jwt_token is not None}")
+        logger.debug("JWT token status", jwt_present=jwt_token is not None)
 
-        print("[DEBUG] Calling langflow_file_service.upload_user_file...")
+        logger.debug("Calling langflow_file_service.upload_user_file")
         result = await langflow_file_service.upload_user_file(
             file_tuple, jwt_token=jwt_token
         )
-        print(f"[DEBUG] Upload successful: {result}")
+        logger.debug("Upload successful", result=result)
         return JSONResponse(result, status_code=201)
     except Exception as e:
-        print(f"[ERROR] upload_user_file endpoint failed: {type(e).__name__}: {e}")
+        logger.error(
+            "upload_user_file endpoint failed",
+            error_type=type(e).__name__,
+            error=str(e),
+        )
         import traceback
 
         print(f"[ERROR] Traceback: {traceback.format_exc()}")
@@ -52,7 +59,8 @@ async def run_ingestion(
         file_ids = payload.get("file_ids")
         file_paths = payload.get("file_paths") or []
         session_id = payload.get("session_id")
-        tweaks = payload.get("tweaks")
+        tweaks = payload.get("tweaks") or {}
+        settings = payload.get("settings", {})
 
         # We assume file_paths is provided. If only file_ids are provided, client would need to resolve to paths via Files API (not implemented here).
         if not file_paths and not file_ids:
@@ -60,6 +68,38 @@ async def run_ingestion(
                 {"error": "Provide file_paths or file_ids"}, status_code=400
             )
 
+        # Convert UI settings to component tweaks using exact component IDs
+        if settings:
+            logger.debug("Applying ingestion settings", settings=settings)
+
+            # Split Text component tweaks (SplitText-QIKhg)
+            if (
+                settings.get("chunkSize")
+                or settings.get("chunkOverlap")
+                or settings.get("separator")
+            ):
+                if "SplitText-QIKhg" not in tweaks:
+                    tweaks["SplitText-QIKhg"] = {}
+                if settings.get("chunkSize"):
+                    tweaks["SplitText-QIKhg"]["chunk_size"] = settings["chunkSize"]
+                if settings.get("chunkOverlap"):
+                    tweaks["SplitText-QIKhg"]["chunk_overlap"] = settings[
+                        "chunkOverlap"
+                    ]
+                if settings.get("separator"):
+                    tweaks["SplitText-QIKhg"]["separator"] = settings["separator"]
+
+            # OpenAI Embeddings component tweaks (OpenAIEmbeddings-joRJ6)
+            if settings.get("embeddingModel"):
+                if "OpenAIEmbeddings-joRJ6" not in tweaks:
+                    tweaks["OpenAIEmbeddings-joRJ6"] = {}
+                tweaks["OpenAIEmbeddings-joRJ6"]["model"] = settings["embeddingModel"]
+
+            # Note: OpenSearch component tweaks not needed for ingestion
+            # (search parameters are for retrieval, not document processing)
+
+            logger.debug("Final tweaks with settings applied", tweaks=tweaks)
+
         # Include user JWT if available
         jwt_token = getattr(request.state, "jwt_token", None)