add Identify errors in plugin

[germankay]

fixes #18

[Copilot]

Pull Request Overview

This PR adds logic to uniquely identify errors in the plugin and prevent duplicate logging and notifications through a new error_logger module.

• Updates the plugin to call log_error with error deduplication and silencing.
• Introduces error_logger.py to handle error logging and notification management.
• Updates the README to document the new error identification and notification features.

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
ckanext/push_errors/plugin.py	Updates error_handler to use log_error with try/except wrapping
ckanext/push_errors/error_logger.py	Adds error_logger module for deduplicated error logging
README.md	Adds documentation for the new error identification and management

Comments suppressed due to low confidence (2)

ckanext/push_errors/error_logger.py:11

• Consider verifying that 'error.traceback' is not None before using it to generate error_id, to avoid always defaulting to 'unknown' when no traceback is available.

def log_error(error, disable_notification=False):

ckanext/push_errors/plugin.py:45

• [nitpick] The variable name 'log_err' is somewhat terse; consider renaming it to something clearer like 'logging_exception' for improved readability.

except Exception as log_err:

[Copilot]

Pull Request Overview

This PR adds error identification enhancements to the plugin by deduplicating error logs and enabling notification suppression, as well as corresponding tests and documentation updates.

• New tests have been added to verify error de-duplication and notification behavior.
• The plugin now uses log_error with error identification and fallback logging.
• The error_logger module implements error deduplication using file path and line number.
• README has been updated to document the new error handling approach.

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated no comments.

File	Description
ckanext/push_errors/tests/test_error_login.py	Adds tests to validate deduplication and optional notification suppression behavior.
ckanext/push_errors/plugin.py	Integrates log_error in the error_handler with fallback logging for log_error failures.
ckanext/push_errors/error_logger.py	Implements error deduplication by tracking errors based on file path and line number.
README.md	Updates documentation for the new error identification and notification management features.

[avdata99]

See comments but not for change, just to analyze. A new PR is required in the latest comment.

[avdata99]

Disable notification is not expected to do like this.
Disable will be dynamic, this PR is to start uniquely identifying error.
File path + line looks good but disable_notification is not required, see other comments

[avdata99]

This will send the push message and after that, we'll sent it again later in this function

[avdata99]

It look interesting this way to get the file_path and the line_number ensure it works, document what else we get (the 3º and 4º params) and use it in the new requested PR

[avdata99]

This will skip all non identified erros. Always a unique ID is required

[avdata99]

This will be always empty.
For the future PR, we need to use redis or the main CKAN DB to preserve a list of logged errors

[avdata99]

We have two type of push-error notifications:

• Server errors captured in make_middleware. This case we have the error object and we are able to traceback
• Log critical messages: we just have a text from the user. We don't have an error object so we can't traceback. Investigate if we can get the last error in python to use in PushErrorHandler.

So the first goal of this PR is to get a unique ID for all messages before sent them. It could be the filepath+line number for error objects and the whole text for log.critical messages.

You can restart a new PR, this code do not capture the general idea.

Let split this into multiple PRs:
Create a new clean PR that:

• Creates a helper function called get_error_trace_id which extracts a unique ID from an error object (something like the file path and line number like this PR already uses)
• Creates a helper function called get_error_log_id which extracts a unique ID from a log records. Use a PushErrorHandler log record as parameter. Investigate is traceback can get the last error in Python. If not, use the message as rthe unique ID. Considering this cuold be long, hash it with some function to reduce it

Use both function and add the unique ID to the push message
Let see how it goes and define the next step