polis-community/red-dwarf

AI Summary: The task is to modify the `red-dwarf` Python library to accept string IDs for participants and statements in the `run_pipeline()` function. Currently, the library assumes numeric IDs, limiting its compatibility. The change requires adapting the code to handle string IDs without breaking existing functionality.

AI Summary: The task is to modify the Red Dwarf project's testing workflow to run tests against multiple Python versions (3.8, 3.10, and 3.13). This involves updating the workflow file to utilize a caching mechanism for different Python versions and ensuring compatibility across these versions. The goal is to improve the project's robustness and compatibility.

AI Summary: The task is to modify the Red Dwarf library's calculation of `group-informed-consensus`. The current calculation incorrectly ranks statements with very low engagement highly. The fix requires adjusting the `p-success` metric to decrease as the number of participating voters decreases, thus lowering the `group-informed-consensus` score for statements with minimal participation. This will ensure that statements with low engagement are ranked appropriately.

AI Summary: Enhance the Red Dwarf library to include a new feature that ranks all statements by their representativeness within each cluster, providing a complete ranked list instead of just the top 5. This involves modifying existing representativeness calculation code to output a comprehensive ranking for each cluster, maintaining the existing functionality while adding the new ranked list.

AI Summary: The task involves enhancing a Python library for dimensional reduction in democratic processes. Specifically, it requires modifying the existing `group-aware-consensus` function to incorporate the probability of disagreement (`pd`) alongside the probability of agreement (`pa`), resulting in a more nuanced understanding of consensus within different groups. The output should be restructured to align with the existing `consensus` type, either as a nested dictionary with 'agree' and 'disagree' keys or a dictionary where each key contains both probabilities.

AI Summary: Document the `data_loader.py` module's API for the Red Dwarf project website. This involves adding docstrings to the `Loader` class and externally used functions, potentially identifying internal functions using underscore prefixes, and updating the project's API reference documentation accordingly.

AI Summary: Implement an export function for the Red Dwarf library that generates data in the same format as Pol.is's export, specifically creating three CSV files: summary, comments, and votes. The function should handle the data structures used internally by Red Dwarf and transform them into the required format for compatibility with Pol.is.

AI Summary: The task is to modify the `red-dwarf` Python library to expose the raw, unfiltered probability data used in its consensus calculation. Currently, the library only returns a filtered subset of the data. The modification should output a complete dataframe of agreement and disagreement probabilities for all statements, either as separate dataframes or a single dataframe with nested 'agree' and 'disagree' fields, as specified in the issue description.