How do I close a duckdb connection

[vzakharov]

When I try to make the second call from my backend (django), I get

  File "/usr/local/lib/python3.11/site-packages/pandasai/smart_dataframe/__init__.py", line 63, in __init__
    self._agent = Agent([df], config=config)
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/pandasai/agent/agent.py", line 28, in __init__
    super().__init__(
  File "/usr/local/lib/python3.11/site-packages/pandasai/agent/base.py", line 75, in __init__
    self.context = PipelineContext(
                   ^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/pandasai/pipelines/pipeline_context.py", line 35, in __init__
    self.cache = cache if cache is not None else Cache()
                                                 ^^^^^^^
  File "/usr/local/lib/python3.11/site-packages/pandasai/helpers/cache.py", line 32, in __init__
    self.connection = duckdb.connect(self.filepath)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
duckdb.duckdb.IOException: IO Error: Could not set lock on file "/app/cache/cache_db_0.11.db": Conflicting lock is held in python (PID 29). See also https://duckdb.org/docs/connect/concurrency

What do I do? Should I close the connection, use a new database cache file every time, else?

[dosubot[bot]]

To close a DuckDB connection in your Django backend and avoid conflicting locks when making a second call, you can use the close method provided in the Cache class. Here is how you can do it:

def close(self) -> None:
    """Close the cache."""
    self.connection.close()

You should call this close method on your Cache instance when you are done with the database operations to ensure the connection is properly closed and avoid any conflicting locks. For example:

# Assuming you have an instance of Cache
cache = Cache()

# Perform your database operations
# ...

# Close the connection when done
cache.close()

This will ensure that the DuckDB connection is properly closed and will help you avoid any conflicting locks when making subsequent calls ^[1].

To continue talking to Dosu, mention @dosu.

[vzakharov]

The fix was to set enable_cache=False when creating the SmartDataframe instance.