File size: 13,529 Bytes
c56a43d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
import os
import json
import requests
import base64
import re
from fastapi import FastAPI
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import SentenceTransformerEmbeddings
from langchain_core.documents import Document

# ───────────────────────────────────────────────
# Configuration
# ───────────────────────────────────────────────
VECTOR_PATH = "./vectorstore/faiss_index"
DOCSTORE_PATH = "./docstore"
FINAL_ANSWER_URL = "https://sameer-handsome173-multi-modal.hf.space/final_answer"
EXTENDED_TIMEOUT = int(os.getenv("FINAL_ANSWER_TIMEOUT", 150))

app = FastAPI(title="πŸ” Multimodal RAG Query Service")

# ───────────────────────────────────────────────
# JSONFileStore
# ───────────────────────────────────────────────
class JSONFileStore:
    def __init__(self, store_path: str):
        self.store_path = store_path
        os.makedirs(self.store_path, exist_ok=True)

    def mget(self, keys: list[str]) -> list[Document]:
        """Retrieve multiple documents by their keys."""
        documents = []
        for key in keys:
            file_path = os.path.join(self.store_path, f"{key}.json")
            if os.path.exists(file_path):
                try:
                    with open(file_path, "r", encoding="utf-8") as f:
                        doc_dict = json.load(f)
                        documents.append(
                            Document(page_content=doc_dict["page_content"], metadata=doc_dict["metadata"])
                        )
                except Exception as e:
                    print(f"Error loading {key}: {e}")
                    documents.append(None)
            else:
                documents.append(None)
        return documents

# ───────────────────────────────────────────────
# Initialize embeddings, vectorstore, docstore
# ───────────────────────────────────────────────
print("πŸ”„ Loading embedding model...")
try:
    embedding_fn = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
    print("βœ… Embedding model loaded")
except Exception as e:
    print(f"❌ Error loading embeddings: {e}")
    raise

try:
    if os.path.exists(VECTOR_PATH):
        vectorstore = FAISS.load_local(VECTOR_PATH, embedding_fn, allow_dangerous_deserialization=True)
        print("βœ… Loaded FAISS vectorstore")
    else:
        raise FileNotFoundError("Vectorstore not found")
except Exception as e:
    print(f"❌ Error loading vectorstore: {e}")
    raise

try:
    if not os.path.exists(DOCSTORE_PATH):
        raise FileNotFoundError("Docstore not found")
    store = JSONFileStore(DOCSTORE_PATH)
    print("βœ… Loaded JSONFileStore")
except Exception as e:
    print(f"❌ Error loading docstore: {e}")
    raise

# ───────────────────────────────────────────────
# Response cleaning helper
# ───────────────────────────────────────────────
def clean_response_text(text: str) -> str:
    """Clean the model's response to remove hashtags, emojis, repetitions and weird tails."""
    if not text:
        return text

    # Remove hashtags and URLs
    text = re.sub(r"#\S+", "", text)
    text = re.sub(r"http\S+", "", text)

    # Remove non-ASCII characters (emojis, special symbols)
    text = text.encode("ascii", "ignore").decode()

    # Remove repeated words sequences (e.g. "word word word")
    text = re.sub(r"\b(\w+)( \1\b)+", r"\1", text, flags=re.IGNORECASE)

    # Collapse multiple newlines and spaces
    text = re.sub(r"\n{2,}", "\n", text)
    text = re.sub(r" {2,}", " ", text).strip()

    # Remove trailing model apology lines or noisy tails
    text = re.sub(r"I'm sorry.*", "", text, flags=re.IGNORECASE)

    return text.strip()

# ───────────────────────────────────────────────
# Helpers for parsing, retrieval and final call
# ───────────────────────────────────────────────
def parse_docs(docs: list[Document]) -> dict:
    """
    Split retrieved documents into images, texts, and tables.
    Returns dict with lists: {"images": [...], "texts": [...], "tables": [...]}
    """
    images, texts, tables = [], [], []

    for doc in docs:
        doc_type = doc.metadata.get("type", "text")
        if doc_type == "image" and doc.metadata.get("is_base64", False):
            # store base64 string
            images.append(doc.page_content)
        elif doc_type == "table":
            tables.append(doc.page_content)
        else:
            texts.append(doc.page_content)

    return {"images": images, "texts": texts, "tables": tables}


def retrieve_documents(query: str, k: int = 5) -> list[Document]:
    """
    Retrieve documents:
    1. Search vectorstore for similar summaries
    2. Collect unique doc_ids from results (avoid duplicates)
    3. Retrieve originals from docstore
    """
    try:
        similar_docs = vectorstore.similarity_search(query, k=k)
        if not similar_docs:
            print("⚠️ No similar documents found")
            return []

        doc_ids = []
        for doc in similar_docs:
            doc_id = doc.metadata.get("doc_id")
            if doc_id and doc_id not in doc_ids:
                doc_ids.append(doc_id)

        if not doc_ids:
            print("⚠️ No doc_ids found in metadata")
            return []

        print(f"πŸ”‘ Found {len(doc_ids)} unique doc_ids")

        original_docs = store.mget(doc_ids)
        original_docs = [d for d in original_docs if d is not None]
        print(f"πŸ“„ Retrieved {len(original_docs)} unique documents")

        return original_docs

    except Exception as e:
        print(f"❌ Error in retrieval: {e}")
        return []


def build_context_and_images(docs_by_type: dict) -> tuple[str, list[str]]:
    """
    Build context text from texts and tables, and collect image base64 strings.
    Returns: (context_text, list_of_base64_images)
    """
    context_parts = []

    # Add text documents
    for i, text_content in enumerate(docs_by_type.get("texts", []), 1):
        context_parts.append(f"--- Text Document {i} ---\n{text_content}")

    # Add table documents
    for i, table_content in enumerate(docs_by_type.get("tables", []), 1):
        context_parts.append(f"--- Table {i} ---\n{table_content}")

    context_text = "\n\n".join(context_parts).strip()
    images_b64 = docs_by_type.get("images", [])

    return context_text, images_b64


def call_final_answer_endpoint(context: str, question: str, images_b64: list[str]) -> dict:
    """
    Call the /final_answer endpoint with context, question, and images.
    Uses extended timeout to allow for slow multimodal inference.
    """
    try:
        # Make prompt instruction clearer for concise output
        data = {
            "context": context,
            "question": f"Answer concisely and without hashtags or emojis.\n\nQuestion: {question}"
        }

        files = []
        if images_b64:
            for i, img_b64 in enumerate(images_b64):
                try:
                    img_bytes = base64.b64decode(img_b64)
                    files.append(("images", (f"image_{i}.jpg", img_bytes, "image/jpeg")))
                except Exception as e:
                    print(f"⚠️ Error decoding image {i}: {e}")

        if files:
            response = requests.post(FINAL_ANSWER_URL, data=data, files=files, timeout=EXTENDED_TIMEOUT)
        else:
            response = requests.post(FINAL_ANSWER_URL, data=data, timeout=EXTENDED_TIMEOUT)

        if response.status_code == 200:
            return response.json()
        else:
            return {"error": f"API returned status {response.status_code}", "details": response.text}

    except Exception as e:
        return {"error": f"Error calling final_answer endpoint: {str(e)}"}

# ───────────────────────────────────────────────
# FastAPI endpoints
# ───────────────────────────────────────────────
@app.get("/")
def home():
    return {
        "message": "βœ… Multimodal RAG Query Service is running",
        "timeout_seconds": EXTENDED_TIMEOUT,
        "endpoints": {
            "query": "/query?question=Your+Question",
            "query_with_details": "/query_with_details?question=Your+Question",
            "stats": "/stats",
        },
    }


@app.get("/stats")
def get_stats():
    try:
        vector_count = vectorstore.index.ntotal if hasattr(vectorstore, "index") else 0
        docstore_files = len([f for f in os.listdir(DOCSTORE_PATH) if f.endswith(".json")]) if os.path.exists(DOCSTORE_PATH) else 0
        return {"status": "ready", "vectorstore_count": vector_count, "docstore_count": docstore_files}
    except Exception as e:
        return {"status": "error", "error": str(e)}


@app.post("/query")
async def query_rag(question: str, k: int = 5):
    """
    Query the Multimodal RAG system:
    1. Search vectorstore for relevant summaries
    2. Retrieve original documents (text + tables + images)
    3. Parse into texts, tables, and images
    4. Call final_answer endpoint with all content
    5. Return cleaned answer
    """
    try:
        print(f"\nπŸ” Query: {question}")

        docs = retrieve_documents(question, k=k)
        if not docs:
            return {"question": question, "answer": "No relevant documents found. Please ingest documents first.", "retrieved_docs": 0}

        docs_by_type = parse_docs(docs)
        print(f"πŸ“Š Parsed: {len(docs_by_type['texts'])} texts, {len(docs_by_type['tables'])} tables, {len(docs_by_type['images'])} images")

        context_text, images_b64 = build_context_and_images(docs_by_type)
        print("πŸš€ Calling final_answer endpoint...")
        result = call_final_answer_endpoint(context_text, question, images_b64)

        if "error" in result:
            return {
                "question": question,
                "error": result["error"],
                "details": result.get("details"),
                "retrieved_docs": len(docs),
                "context_preview": context_text[:300] if context_text else "No context"
            }

        cleaned_answer = clean_response_text(result.get("response", "No response generated"))

        return {
            "question": question,
            "answer": cleaned_answer,
            "retrieved_docs": len(docs),
            "docs_info": {
                "texts": len(docs_by_type["texts"]),
                "tables": len(docs_by_type["tables"]),
                "images": len(docs_by_type["images"]),
            },
            "context_preview": context_text[:300] if context_text else "No context",
        }

    except Exception as e:
        import traceback
        return {"question": question, "error": str(e), "traceback": traceback.format_exc()}


@app.post("/query_with_details")
async def query_with_details(question: str, k: int = 5):
    """Query with detailed document information"""
    try:
        print(f"\nπŸ” Detailed Query: {question}")

        docs = retrieve_documents(question, k=k)
        if not docs:
            return {"question": question, "answer": "No relevant documents found.", "retrieved_docs": []}

        docs_by_type = parse_docs(docs)
        context_text, images_b64 = build_context_and_images(docs_by_type)

        result = call_final_answer_endpoint(context_text, question, images_b64)

        if "error" in result:
            return {"question": question, "error": result["error"], "details": result.get("details")}

        docs_info = []
        for doc in docs:
            doc_info = {
                "doc_id": doc.metadata.get("doc_id"),
                "type": doc.metadata.get("type"),
                "source": doc.metadata.get("source"),
                "summary": doc.metadata.get("summary", "")[:200],
            }
            doc_info["content"] = "[Base64 Image Data]" if doc.metadata.get("type") == "image" else doc.page_content[:300]
            docs_info.append(doc_info)

        cleaned_answer = clean_response_text(result.get("response", "No response generated"))

        return {
            "question": question,
            "answer": cleaned_answer,
            "retrieved_docs": docs_info,
            "stats": {
                "total_retrieved": len(docs),
                "texts": len(docs_by_type["texts"]),
                "tables": len(docs_by_type["tables"]),
                "images": len(docs_by_type["images"]),
            },
        }

    except Exception as e:
        import traceback
        return {"error": str(e), "traceback": traceback.format_exc()}