Refactor: Drop MongoDB (#48)

* feat: replace mongodb with sqllite

* feat: update docker compose to drop mongo

* chore: drop logs

* chore: cleanup

* fix: unit tests

* fix: workflow

* fix: workflow run

.github/workflows/docker-image.yml

@@ -7,7 +7,7 @@ on:
 
 jobs:
   build:
-    if: ${{ github.event.workflow_run.conclusion == 'success' && github.ref == 'refs/heads/master' && github.event_name != 'pull_request' }}
+    if: ${{ github.event.workflow_run.conclusion == 'success' && github.event.workflow_run.head_branch == 'master' }}
     runs-on: ubuntu-latest
     steps:
       - name: Checkout

.github/workflows/unit-tests.yml

@@ -15,6 +15,9 @@ jobs:
       - name: Checkout
         uses: actions/checkout@v4
 
+      - name: Set env
+        run: echo "ENV=test" >> $GITHUB_ENV
+
       - name: Install pdm
         run: pip install pdm
 

.gitignore

@@ -187,3 +187,4 @@ cython_debug/
 postgres_data
 .vscode
 ollama
+data

api/backend/app.py

@@ -13,6 +13,7 @@
 from api.backend.routers.job_router import job_router
 from api.backend.routers.log_router import log_router
 from api.backend.routers.stats_router import stats_router
+from api.backend.database.startup import init_database
 
 log_level = os.getenv("LOG_LEVEL")
 LOG_LEVEL = get_log_level(log_level)
@@ -41,3 +42,10 @@
 app.include_router(job_router)
 app.include_router(log_router)
 app.include_router(stats_router)
+
+
+@app.on_event("startup")
+async def startup_event():
+    if os.getenv("ENV") != "test":
+        init_database()
+        LOG.info("Starting up...")

api/backend/auth/auth_router.py

@@ -7,7 +7,6 @@
 
 # LOCAL
 from api.backend.schemas import User, Token, UserCreate
-from api.backend.database import get_user_collection
 from api.backend.auth.auth_utils import (
     ACCESS_TOKEN_EXPIRE_MINUTES,
     get_current_user,
@@ -16,6 +15,8 @@
     create_access_token,
 )
 
+from api.backend.database.common import update
+
 auth_router = APIRouter()
 
 
@@ -43,12 +44,13 @@ async def login_for_access_token(form_data: OAuth2PasswordRequestForm = Depends(
 
 @auth_router.post("/auth/signup", response_model=User)
 async def create_user(user: UserCreate):
-    users_collection = get_user_collection()
     hashed_password = get_password_hash(user.password)
     user_dict = user.model_dump()
     user_dict["hashed_password"] = hashed_password
     del user_dict["password"]
-    _ = await users_collection.insert_one(user_dict)
+
+    query = "INSERT INTO users (email, hashed_password, full_name) VALUES (?, ?, ?)"
+    _ = update(query, (user_dict["email"], hashed_password, user_dict["full_name"]))
     return user_dict
 
 

api/backend/auth/auth_utils.py

@@ -15,7 +15,8 @@
 
 # LOCAL
 from api.backend.schemas import User, UserInDB, TokenData
-from api.backend.database import get_user_collection
+
+from api.backend.database.common import query
 
 LOG = logging.getLogger(__name__)
 
@@ -40,8 +41,8 @@ def get_password_hash(password: str):
 
 
 async def get_user(email: str):
-    user_collection = get_user_collection()
-    user = await user_collection.find_one({"email": email})
+    user_query = "SELECT * FROM users WHERE email = ?"
+    user = query(user_query, (email,))[0]
 
     if not user:
         return

api/backend/constants.py

@@ -0,0 +1 @@
+DATABASE_PATH = "data/database.db"

api/backend/database/__init__.py

@@ -0,0 +1,3 @@
+from .common import insert, QUERIES, update
+
+__all__ = ["insert", "QUERIES", "update"]

api/backend/database/common.py

@@ -0,0 +1,92 @@
+import sqlite3
+from typing import Any, Optional
+from api.backend.constants import DATABASE_PATH
+from api.backend.utils import format_json, format_sql_row_to_python
+from api.backend.database.schema import INIT_QUERY
+from api.backend.database.queries import JOB_INSERT_QUERY, DELETE_JOB_QUERY
+import logging
+
+LOG = logging.getLogger(__name__)
+
+
+def connect():
+    connection = sqlite3.connect(DATABASE_PATH)
+    connection.set_trace_callback(print)
+    cursor = connection.cursor()
+    return cursor
+
+
+def insert(query: str, values: tuple[Any, ...]):
+    connection = sqlite3.connect(DATABASE_PATH)
+    cursor = connection.cursor()
+    copy = list(values)
+    format_json(copy)
+
+    try:
+        _ = cursor.execute(query, copy)
+        connection.commit()
+    except sqlite3.Error as e:
+        LOG.error(f"An error occurred: {e}")
+    finally:
+        cursor.close()
+        connection.close()
+
+
+def query(query: str, values: Optional[tuple[Any, ...]] = None):
+    connection = sqlite3.connect(DATABASE_PATH)
+    connection.row_factory = sqlite3.Row
+    cursor = connection.cursor()
+    rows = []
+    try:
+        if values:
+            _ = cursor.execute(query, values)
+        else:
+            _ = cursor.execute(query)
+
+        rows = cursor.fetchall()
+
+    finally:
+        cursor.close()
+        connection.close()
+
+    formatted_rows: list[dict[str, Any]] = []
+
+    for row in rows:
+        row = dict(row)
+        formatted_row = format_sql_row_to_python(row)
+        formatted_rows.append(formatted_row)
+
+    return formatted_rows
+
+
+def update(query: str, values: Optional[tuple[Any, ...]] = None):
+    connection = sqlite3.connect(DATABASE_PATH)
+    cursor = connection.cursor()
+
+    copy = None
+
+    if values:
+        copy = list(values)
+        format_json(copy)
+
+    try:
+        if copy:
+            res = cursor.execute(query, copy)
+        else:
+            res = cursor.execute(query)
+        connection.commit()
+        return res.rowcount
+    except sqlite3.Error as e:
+        LOG.error(f"An error occurred: {e}")
+    finally:
+        cursor.close()
+        connection.close()
+
+    return 0
+
+
+QUERIES = {
+    "init": INIT_QUERY,
+    "insert_job": JOB_INSERT_QUERY,
+    "delete_job": DELETE_JOB_QUERY,
+}

api/backend/database/queries/__init__.py

@@ -0,0 +1,3 @@
+from .queries import JOB_INSERT_QUERY, DELETE_JOB_QUERY
+
+__all__ = ["JOB_INSERT_QUERY", "DELETE_JOB_QUERY"]

api/backend/database/queries/queries.py

@@ -0,0 +1,9 @@
+JOB_INSERT_QUERY = """
+INSERT INTO jobs 
+(id, url, elements, user, time_created, result, status, chat, job_options)
+VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?)
+"""
+
+DELETE_JOB_QUERY = """
+DELETE FROM jobs WHERE id IN ()
+"""

api/backend/database/schema/__init__.py

@@ -0,0 +1,3 @@
+from .schema import INIT_QUERY
+
+__all__ = ["INIT_QUERY"]

api/backend/database/schema/schema.py

@@ -0,0 +1,20 @@
+INIT_QUERY = """
+CREATE TABLE IF NOT EXISTS jobs (
+    id STRING PRIMARY KEY NOT NULL,
+    url STRING NOT NULL,
+    elements JSON NOT NULL,
+    user STRING,
+    time_created DATETIME NOT NULL,
+    result JSON NOT NULL,
+    status STRING NOT NULL,
+    chat JSON,
+    job_options JSON
+);
+
+CREATE TABLE IF NOT EXISTS users (
+    email STRING PRIMARY KEY NOT NULL,
+    hashed_password STRING NOT NULL,
+    full_name STRING,
+    disabled BOOLEAN
+);
+"""

api/backend/database/startup.py

@@ -0,0 +1,15 @@
+from api.backend.database.common import connect, QUERIES
+import logging
+
+LOG = logging.getLogger(__name__)
+
+
+def init_database():
+    cursor = connect()
+
+    for query in QUERIES["init"].strip().split(";"):
+        if query.strip():
+            LOG.info(f"Executing query: {query}")
+            _ = cursor.execute(query)
+
+    cursor.close()

api/backend/job/__init__.py

@@ -1,5 +1,4 @@
 from .job import (
-    query,
     insert,
     update_job,
     delete_jobs,
@@ -9,7 +8,6 @@
 )
 
 __all__ = [
-    "query",
     "insert",
     "update_job",
     "delete_jobs",