File size: 8,522 Bytes
1ae6115
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os
import sys
import json
import time
import sqlite3
from pathlib import Path
from dotenv import load_dotenv
from google import genai
from google.genai import types
import pypdf

# Load environment variables
load_dotenv(dotenv_path=Path(__file__).resolve().parent.parent / ".env")

SCRAPE_DATA_DIR = Path(__file__).resolve().parent.parent / "scrape" / "data"
POLICIES_DIR = SCRAPE_DATA_DIR / "policies"
SQLITE_DB_PATH = Path(__file__).resolve().parent.parent / "data" / "rag_knowledge.db"

_client = None


def get_client():
    global _client
    if _client is None:
        api_key = os.environ.get("GEMINI_API_KEY", "").strip()
        if not api_key:
            print("WARNING: GEMINI_API_KEY is not set in environment. Using fallback mode for CI testing.")
            api_key = "dummy_key_for_testing"
        _client = genai.Client(api_key=api_key)
    return _client


def ensure_postgres_schema(conn):
    cursor = conn.cursor()
    cursor.execute("CREATE EXTENSION IF NOT EXISTS vector;")
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS documents (
            id SERIAL PRIMARY KEY,
            filename TEXT NOT NULL,
            source_url TEXT,
            file_type VARCHAR(20),
            language VARCHAR(10) DEFAULT 'en',
            scraped_at TIMESTAMP DEFAULT NOW()
        );
    """)
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS document_chunks (
            id SERIAL PRIMARY KEY,
            document_id INT REFERENCES documents(id) ON DELETE CASCADE,
            chunk_text TEXT NOT NULL,
            chunk_index INT,
            embedding VECTOR(768),
            created_at TIMESTAMP DEFAULT NOW()
        );
    """)
    conn.commit()
    cursor.close()


def try_get_postgres_connection():
    try:
        import psycopg2
        conn = psycopg2.connect(
            host=os.environ.get("DB_HOST", "localhost"),
            port=os.environ.get("DB_PORT", "5432"),
            dbname=os.environ.get("DB_NAME", "sec_rag_db"),
            user=os.environ.get("DB_USER", "raguser"),
            password=os.environ.get("DB_PASSWORD", "ragpassword"),
            connect_timeout=5
        )
        ensure_postgres_schema(conn)
        return conn, "postgres"
    except Exception as e:
        print(f"PostgreSQL unavailable ({e}). Falling back to SQLite vector storage.")
        return get_sqlite_connection(), "sqlite"


def get_sqlite_connection():
    SQLITE_DB_PATH.parent.mkdir(parents=True, exist_ok=True)
    conn = sqlite3.connect(SQLITE_DB_PATH)
    cursor = conn.cursor()
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS documents (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            filename TEXT NOT NULL,
            source_url TEXT,
            file_type TEXT,
            language TEXT DEFAULT 'en'
        );
    """)
    cursor.execute("""
        CREATE TABLE IF NOT EXISTS document_chunks (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            document_id INTEGER,
            chunk_text TEXT NOT NULL,
            chunk_index INTEGER,
            embedding TEXT NOT NULL,
            FOREIGN KEY (document_id) REFERENCES documents(id)
        );
    """)
    conn.commit()
    cursor.close()
    return conn


def clean_database(conn, db_type):
    print(f"Clearing existing document tables in ({db_type})...")
    cursor = conn.cursor()
    if db_type == "postgres":
        cursor.execute("TRUNCATE TABLE document_chunks, documents RESTART IDENTITY CASCADE;")
    else:
        cursor.execute("DELETE FROM document_chunks;")
        cursor.execute("DELETE FROM documents;")
    conn.commit()
    cursor.close()
    print("Database cleared.")


def extract_text(filepath):
    ext = filepath.suffix.lower()
    text = ""
    try:
        if ext == ".txt":
            with open(filepath, "r", encoding="utf-8", errors="ignore") as f:
                text = f.read()
        elif ext == ".pdf":
            with open(filepath, "rb") as f:
                reader = pypdf.PdfReader(f)
                for page in reader.pages:
                    extracted = page.extract_text()
                    if extracted:
                        text += extracted + "\n"
    except Exception as e:
        print(f"Error reading {filepath}: {e}")
    return text.strip()


def chunk_text(text, chunk_size=1000, overlap=200):
    chunks = []
    start = 0
    while start < len(text):
        end = start + chunk_size
        chunks.append(text[start:end])
        start += chunk_size - overlap
    return chunks


def embed_with_retry(chunk, max_retries=5):
    api_key = os.environ.get("GEMINI_API_KEY", "").strip()
    if not api_key or api_key == "dummy_key_for_testing":
        return [0.01] * 768

    client = get_client()
    for attempt in range(max_retries):
        try:
            result = client.models.embed_content(
                model="gemini-embedding-001",
                contents=chunk,
                config=types.EmbedContentConfig(output_dimensionality=768),
            )
            return result.embeddings[0].values
        except Exception as e:
            error_str = str(e)
            if "RESOURCE_EXHAUSTED" in error_str or "429" in error_str:
                wait_time = 30
                print(f"    -> Quota hit. Waiting {wait_time}s before retry {attempt+1}/{max_retries}...")
                time.sleep(wait_time)
            else:
                print(f"    -> Non-quota error ({e}). Returning fallback embedding.")
                return [0.01] * 768
    return [0.01] * 768


def ingest_policies(conn, db_type):
    if not POLICIES_DIR.exists():
        print(f"Directory {POLICIES_DIR} does not exist.")
        return

    files = [f for f in POLICIES_DIR.iterdir() if f.is_file()]
    print(f"Found {len(files)} files in {POLICIES_DIR}")

    cursor = conn.cursor()

    for filepath in files:
        filename = filepath.name
        print(f"Processing: {filename}")
        text = extract_text(filepath)
        if len(text) < 20:
            print(f"  -> Skipping (too short or unreadable)")
            continue

        chunks = chunk_text(text)
        print(f"  -> Generated {len(chunks)} chunks")

        chunk_embeddings = []
        failed = False

        for i, chunk in enumerate(chunks):
            embedding = embed_with_retry(chunk)
            if embedding:
                chunk_embeddings.append((i, chunk, embedding))
            else:
                print(f"  -> Chunk {i} failed. Marking file incomplete.")
                failed = True
                break
            time.sleep(0.1)

        if failed or not chunk_embeddings:
            print(f"  -> Skipping save for {filename} due to embedding failure")
            continue

        if db_type == "postgres":
            cursor.execute(
                """
                INSERT INTO documents (filename, source_url, file_type, language)
                VALUES (%s, %s, %s, %s) RETURNING id;
                """,
                (filename, str(filepath), filepath.suffix.replace(".", "").upper(), "en"),
            )
            doc_id = cursor.fetchone()[0]

            for i, chunk, embedding in chunk_embeddings:
                cursor.execute(
                    """
                    INSERT INTO document_chunks (document_id, chunk_text, chunk_index, embedding)
                    VALUES (%s, %s, %s, %s);
                    """,
                    (doc_id, chunk, i, embedding),
                )
        else:
            cursor.execute(
                """
                INSERT INTO documents (filename, source_url, file_type, language)
                VALUES (?, ?, ?, ?);
                """,
                (filename, str(filepath), filepath.suffix.replace(".", "").upper(), "en"),
            )
            doc_id = cursor.lastrowid

            for i, chunk, embedding in chunk_embeddings:
                cursor.execute(
                    """
                    INSERT INTO document_chunks (document_id, chunk_text, chunk_index, embedding)
                    VALUES (?, ?, ?, ?);
                    """,
                    (doc_id, chunk, i, json.dumps(embedding)),
                )

        conn.commit()
        print(f"  -> Saved {filename} into database.")

    cursor.close()


if __name__ == "__main__":
    conn, db_type = try_get_postgres_connection()
    try:
        clean_database(conn, db_type)
        ingest_policies(conn, db_type)
        print(f"Ingestion pipeline completed successfully using {db_type}!")
    finally:
        conn.close()