fix

.env

@@ -15,4 +15,9 @@ CHROMA_DIR=C:/Users/DELL/Desktop/Deploymnet/29 oct/py-learn-backend/ragg/chroma
 CHROMA_ROOT=C:/Users/DELL/Desktop/Deploymnet/29 oct/py-learn-backend/ragg/chroma
 EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2
 ALLOWED_ORIGINS=http://localhost:4200,http://127.0.0.1:4200
-RAG_INGEST_URL=http://localhost:5000/rag/ingest
+RAG_INGEST_URL=http://localhost:5000/rag/ingest
+AWS_ACCESS_KEY_ID=AKIA3PWGNRHL7RTV3XRJ
+AWS_SECRET_ACCESS_KEY=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
+AWS_REGION=ap-south-1
+S3_BUCKET=pykara-tts-audio
+S3_PREFIX=audio/

media/audio/explain_0f06635a143f44bd9e705a53485476d7.wav

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:189a4ca997451cafe72c7652a89ccd0f43fc107064a075624af5499b10f219fe
-size 1092844

media/audio/explain_1bf217407423408f8800bcb2a3db2d3b.wav

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:c0f7690ce19971fba91c94c34076908f61d335d88ddbcf456d57a98c53fd155b
-size 1305324

ragg/app.py

@@ -12,6 +12,13 @@ from flask_cors import CORS
 import requests
 from TTS.api import TTS
 
+# --- S3 (added) ---
+try:
+    import boto3
+    from botocore.exceptions import NoCredentialsError, ClientError
+except Exception:
+    boto3 = None
+    NoCredentialsError = ClientError = Exception  # fallbacks so type names exist
 
 # RAG imports
 try:
@@ -103,6 +110,49 @@ try:
 except Exception:
     pass
 
+# --- S3 config (added) ---
+S3_BUCKET = os.getenv("S3_BUCKET", "").strip()
+AWS_REGION = os.getenv("AWS_REGION", "ap-south-1").strip()
+S3_PREFIX = os.getenv("S3_PREFIX", "audio/").strip()
+AWS_ACCESS_KEY_ID = os.getenv("AWS_ACCESS_KEY_ID", "").strip()
+AWS_SECRET_ACCESS_KEY = os.getenv("AWS_SECRET_ACCESS_KEY", "").strip()
+
+_s3_client = None
+if boto3 and S3_BUCKET and AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY:
+    try:
+        _s3_client = boto3.client(
+            "s3",
+            region_name=AWS_REGION,
+            aws_access_key_id=AWS_ACCESS_KEY_ID,
+            aws_secret_access_key=AWS_SECRET_ACCESS_KEY,
+        )
+    except Exception as _e:
+        _s3_client = None
+
+def _upload_to_s3(file_path: Union[str, Path]) -> Optional[str]:
+    """
+    Upload the file to S3 and return a presigned URL (24h).
+    If S3 is not configured, returns None (caller will fallback).
+    """
+    if not _s3_client or not S3_BUCKET:
+        return None
+    try:
+        file_path = str(file_path)
+        key = f"{S3_PREFIX}{Path(file_path).name}"
+        _s3_client.upload_file(file_path, S3_BUCKET, key)
+        url = _s3_client.generate_presigned_url(
+            "get_object",
+            Params={"Bucket": S3_BUCKET, "Key": key},
+            ExpiresIn=24 * 3600,
+        )
+        return url
+    except (NoCredentialsError, ClientError) as e:
+        try:
+            current_app.logger.error(f"S3 upload failed: {e}")
+        except Exception:
+            print(f"S3 upload failed: {e}")
+        return None
+
 # Media and voice references
 
 # MEDIA_ROOT = Path(os.getenv("MEDIA_ROOT", "./media"))
@@ -320,14 +370,18 @@ def rag_explain_grammar():
                     reference_files=None,
                     language=data.get("language", "en"),
                 )
-                # ✅ Fixed for Hugging Face audio URL
+                # Local: serve from /rag/audio/*
                 if "localhost" in request.host_url or "127.0.0.1" in request.host_url:
                     base = request.host_url.rstrip("/")
+                    result_dict["audio_url"] = f"{base}/rag/audio/{wav_path.name}"
                 else:
-                    base = os.getenv("SPACE_URL", "https://pykara-py-learn-backend.hf.space")
-                result_dict["audio_url"] = f"{base}/rag/audio/{wav_path.name}"
-                # base = request.host_url.rstrip("/")
-                # result_dict["audio_url"] = f"{base}/rag/audio/{wav_path.name}"
+                    # Deployed: try S3 first; fallback to public SPACE_URL if set
+                    s3_url = _upload_to_s3(str(wav_path))
+                    if s3_url:
+                        result_dict["audio_url"] = s3_url
+                    else:
+                        base = os.getenv("SPACE_URL", "https://pykara-py-learn-backend.hf.space")
+                        result_dict["audio_url"] = f"{base}/rag/audio/{wav_path.name}"
             except FileNotFoundError as e:
                 current_app.logger.error("XTTS reference audio missing: %s", e)
             except Exception as e:
@@ -582,15 +636,19 @@ def rag_synthesize_audio():
             reference_files=reference_files,
             language=language,
         )
-        # ✅ Fixed for Hugging Face audio URL
+        # Local: serve static file
         if "localhost" in request.host_url or "127.0.0.1" in request.host_url:
             base = request.host_url.rstrip("/")
+            audio_url = f"{base}/rag/audio/{wav_path.name}"
         else:
-            base = os.getenv("SPACE_URL", "https://pykara-py-learn-backend.hf.space")
-        audio_url = f"{base}/rag/audio/{wav_path.name}"
+            # Deployed: try S3 first; fallback to SPACE_URL
+            s3_url = _upload_to_s3(str(wav_path))
+            if s3_url:
+                audio_url = s3_url
+            else:
+                base = os.getenv("SPACE_URL", "https://pykara-py-learn-backend.hf.space")
+                audio_url = f"{base}/rag/audio/{wav_path.name}"
 
-        # base = request.host_url.rstrip("/")  # e.g., http://localhost:7000
-        # audio_url = f"{base}/rag/audio/{wav_path.name}"
         return jsonify({"audio_url": audio_url, "file": wav_path.name}), 200
     except FileNotFoundError as e:
         current_app.logger.error("XTTS references missing: %s", e)