fix: merge

Author: suryaiyer95

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.0.19
+current_version = 0.0.20
 commit = True
 tag = True
 

docs/conf.py

@@ -15,7 +15,7 @@
 year = "2024"
 author = "Altimate Inc."
 copyright = f"{year}, {author}"
-version = release = "0.0.19"
+version = release = "0.0.20"
 
 pygments_style = "trac"
 templates_path = ["."]

setup.py

@@ -13,7 +13,7 @@ def read(*names, **kwargs):
 
 setup(
     name="altimate-datapilot-cli",
-    version="0.0.19",
+    version="0.0.20",
     license="MIT",
     description="Assistant for Data Teams",
     long_description="{}\n{}".format(
@@ -70,6 +70,7 @@ def read(*names, **kwargs):
         "sqlglot~=25.30.0",
         "mcp~=1.9.0",
         "pyperclip~=1.8.2",
+        "python-dotenv~=1.0.0",
     ],
     extras_require={
         # eg:

src/datapilot/__init__.py

@@ -1 +1 @@
-__version__ = "0.0.19"
+__version__ = "0.0.20"

src/datapilot/cli/main.py

@@ -1,13 +1,96 @@
+import json
+import os
+import re
+from pathlib import Path
+
 import click
+from dotenv import load_dotenv
 
+from datapilot import __version__
+from datapilot.core.knowledge.cli import cli as knowledge
 from datapilot.core.mcp_utils.mcp import mcp
 from datapilot.core.platforms.dbt.cli.cli import dbt
 
 
+def load_config_from_file():
+    """Load configuration from ~/.altimate/altimate.json if it exists."""
+    config_path = Path.home() / ".altimate" / "altimate.json"
+
+    if not config_path.exists():
+        return {}
+
+    try:
+        with config_path.open() as f:
+            config = json.load(f)
+        return config
+    except (OSError, json.JSONDecodeError) as e:
+        click.echo(f"Warning: Failed to load config from {config_path}: {e}", err=True)
+        return {}
+
+
+def substitute_env_vars(value):
+    """Replace ${env:ENV_VARIABLE} patterns with actual environment variable values."""
+    if not isinstance(value, str):
+        return value
+
+    # Pattern to match ${env:VARIABLE_NAME}
+    pattern = r"\$\{env:([^}]+)\}"
+
+    def replacer(match):
+        env_var = match.group(1)
+        return os.environ.get(env_var, match.group(0))
+
+    return re.sub(pattern, replacer, value)
+
+
+def process_config(config):
+    """Process configuration dictionary to substitute environment variables."""
+    processed = {}
+    for key, value in config.items():
+        processed[key] = substitute_env_vars(value)
+    return processed
+
+
 @click.group()
-def datapilot():
+@click.version_option(version=__version__, prog_name="datapilot")
+@click.option("--token", required=False, help="Your API token for authentication.", hide_input=True)
+@click.option("--instance-name", required=False, help="Your tenant ID.")
+@click.option("--backend-url", required=False, help="Altimate's Backend URL", default="https://api.myaltimate.com")
+@click.pass_context
+def datapilot(ctx, token, instance_name, backend_url):
     """Altimate CLI for DBT project management."""
+    # Load .env file from current directory if it exists
+    load_dotenv()
+
+    # Load configuration from file
+    file_config = load_config_from_file()
+    file_config = process_config(file_config)
+
+    # Map config file keys to CLI option names
+    config_mapping = {"altimateApiKey": "token", "altimateInstanceName": "instance_name", "altimateUrl": "backend_url"}
+
+    # Store common options in context, with CLI args taking precedence
+    ctx.ensure_object(dict)
+
+    # Apply file config first
+    for file_key, cli_key in config_mapping.items():
+        if file_key in file_config:
+            ctx.obj[cli_key] = file_config[file_key]
+
+    # Override with CLI arguments if provided
+    if token is not None:
+        ctx.obj["token"] = token
+    if instance_name is not None:
+        ctx.obj["instance_name"] = instance_name
+    if backend_url != "https://api.myaltimate.com":  # Only override if not default
+        ctx.obj["backend_url"] = backend_url
+
+    # Set defaults if nothing was provided
+    ctx.obj.setdefault("token", None)
+    ctx.obj.setdefault("instance_name", None)
+    ctx.obj.setdefault("backend_url", "https://api.myaltimate.com")
 
 
 datapilot.add_command(dbt)
 datapilot.add_command(mcp)
+datapilot.add_command(knowledge)

src/datapilot/core/knowledge/__init__.py

@@ -0,0 +1,46 @@
+from http.server import HTTPServer
+
+import click
+
+from .server import KnowledgeBaseHandler
+
+
+@click.group(name="knowledge")
+def cli():
+    """knowledge specific commands."""
+
+
+@cli.command()
+@click.option("--port", default=4000, help="Port to run the server on")
+@click.pass_context
+def serve(ctx, port):
+    """Serve knowledge bases via HTTP server."""
+    # Get configuration from parent context
+    token = ctx.parent.obj.get("token")
+    instance_name = ctx.parent.obj.get("instance_name")
+    backend_url = ctx.parent.obj.get("backend_url")
+
+    if not token or not instance_name:
+        click.echo(
+            "Error: API token and instance name are required. Use --token and --instance-name options or set them in config.", err=True
+        )
+        ctx.exit(1)
+
+    # Set context data for the handler
+    KnowledgeBaseHandler.token = token
+    KnowledgeBaseHandler.instance_name = instance_name
+    KnowledgeBaseHandler.backend_url = backend_url
+
+    server_address = ("", port)
+    httpd = HTTPServer(server_address, KnowledgeBaseHandler)
+
+    click.echo(f"Starting knowledge base server on port {port}...")
+    click.echo(f"Backend URL: {backend_url}")
+    click.echo(f"Instance: {instance_name}")
+    click.echo(f"Server running at http://localhost:{port}")
+
+    try:
+        httpd.serve_forever()
+    except KeyboardInterrupt:
+        click.echo("\nShutting down server...")
+        httpd.shutdown()

src/datapilot/core/knowledge/cli.py

@@ -0,0 +1,84 @@
+import json
+import re
+from http.server import BaseHTTPRequestHandler
+from urllib.error import HTTPError
+from urllib.error import URLError
+from urllib.parse import urlparse
+from urllib.request import Request
+from urllib.request import urlopen
+
+import click
+
+
+class KnowledgeBaseHandler(BaseHTTPRequestHandler):
+    """HTTP request handler for serving knowledge bases and health checks."""
+
+    token: str = ""
+    instance_name: str = ""
+    backend_url: str = ""
+
+    def do_GET(self):
+        """Handle GET requests."""
+        path = urlparse(self.path).path
+
+        # Match /knowledge_bases/{uuid} pattern
+        match = re.match(r"^/kb/([a-fA-F0-9]{8}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{12})$", path)
+
+        if match:
+            public_id = match.group(1)
+            self.handle_knowledge_base(public_id)
+        elif path == "/health":
+            self.handle_health()
+        else:
+            self.send_error(404, "Not Found")
+
+    def handle_knowledge_base(self, public_id):
+        """Fetch and return knowledge base data."""
+        url = f"{self.backend_url}/knowledge_bases/private/{public_id}"
+
+        # Validate URL scheme for security
+        parsed_url = urlparse(url)
+        if parsed_url.scheme not in ("http", "https"):
+            self.send_response(400)
+            self.send_header("Content-Type", "application/json")
+            self.end_headers()
+            error_msg = json.dumps({"error": "Invalid URL scheme. Only HTTP and HTTPS are allowed."})
+            self.wfile.write(error_msg.encode("utf-8"))
+            return
+
+        headers = {"Authorization": f"Bearer {self.token}", "X-Tenant": self.instance_name, "Content-Type": "application/json"}
+
+        req = Request(url, headers=headers)  # noqa: S310
+
+        try:
+            # URL scheme validated above - only HTTP/HTTPS allowed
+            with urlopen(req, timeout=30) as response:  # noqa: S310
+                data = response.read()
+                self.send_response(200)
+                self.send_header("Content-Type", "application/json")
+                self.end_headers()
+                self.wfile.write(data)
+        except HTTPError as e:
+            error_body = e.read()
+            error_data = error_body.decode("utf-8") if error_body else '{"error": "HTTP Error"}'
+            self.send_response(e.code)
+            self.send_header("Content-Type", "application/json")
+            self.end_headers()
+            self.wfile.write(error_data.encode("utf-8"))
+        except URLError as e:
+            self.send_response(500)
+            self.send_header("Content-Type", "application/json")
+            self.end_headers()
+            error_msg = json.dumps({"error": str(e)})
+            self.wfile.write(error_msg.encode("utf-8"))
+
+    def handle_health(self):
+        """Handle health check endpoint."""
+        self.send_response(200)
+        self.send_header("Content-Type", "application/json")
+        self.end_headers()
+        self.wfile.write(json.dumps({"status": "ok"}).encode("utf-8"))
+
+    def log_message(self, format, *args):
+        """Override to use click.echo for logging."""
+        click.echo(f"{self.address_string()} - {format % args}")

src/datapilot/core/knowledge/server.py

@@ -24,13 +24,14 @@
 
 # New dbt group
 @click.group()
-def dbt():
+@click.pass_context
+def dbt(ctx):
     """DBT specific commands."""
+    # Ensure context object exists
+    ctx.ensure_object(dict)
 
 
 @dbt.command("project-health")
-@click.option("--token", required=False, help="Your API token for authentication.")
-@click.option("--instance-name", required=False, help="Your tenant ID.")
 @click.option(
     "--manifest-path",
     required=True,
@@ -57,21 +58,24 @@ def dbt():
     default=None,
     help="Selective model testing. Specify one or more models to run tests on.",
 )
-@click.option("--backend-url", required=False, help="Altimate's Backend URL", default="https://api.myaltimate.com")
+@click.pass_context
 def project_health(
-    token,
-    instance_name,
+    ctx,
     manifest_path,
     catalog_path,
     config_path=None,
     config_name=None,
     select=None,
-    backend_url="https://api.myaltimate.com",
 ):
     """
     Validate the DBT project's configuration and structure.
     :param manifest_path: Path to the DBT manifest file.
     """
+    # Get common options from parent context
+    token = ctx.parent.obj.get("token")
+    instance_name = ctx.parent.obj.get("instance_name")
+    backend_url = ctx.parent.obj.get("backend_url")
+
     config = None
     if config_path:
         config = load_config(config_path)
@@ -131,8 +135,7 @@ def project_health(
 
 
 @dbt.command("onboard")
-@click.option("--token", prompt="API Token", help="Your API token for authentication.")
-@click.option("--instance-name", prompt="Instance Name", help="Your tenant ID.")
+@click.option("--dbt_core_integration_id", prompt="DBT Core Integration ID", help="DBT Core Integration ID")
 @click.option(
     "--dbt_core_integration_id",
     "--dbt_integration_id",
@@ -150,17 +153,26 @@ def project_health(
 )
 @click.option("--manifest-path", required=True, prompt="Manifest Path", help="Path to the manifest file.")
 @click.option("--catalog-path", required=False, prompt=False, help="Path to the catalog file.")
-@click.option("--backend-url", required=False, help="Altimate's Backend URL", default="https://api.myaltimate.com")
+@click.pass_context
 def onboard(
-    token,
-    instance_name,
-    dbt_integration_id,  # Updated parameter name
-    dbt_integration_environment,  # Updated parameter name
+    ctx,
+    dbt_integration_id,
+    dbt_integration_environment,
     manifest_path,
     catalog_path,
-    backend_url="https://api.myaltimate.com",
 ):
     """Onboard a manifest file to DBT."""
+    # Get common options from parent context
+    token = ctx.parent.obj.get("token")
+    instance_name = ctx.parent.obj.get("instance_name")
+    backend_url = ctx.parent.obj.get("backend_url")
+
+    # For onboard command, token and instance_name are required
+    if not token:
+        token = click.prompt("API Token")
+    if not instance_name:
+        instance_name = click.prompt("Instance Name")
+
     check_token_and_instance(token, instance_name)
 
     if not validate_credentials(token, backend_url, instance_name):