[BUG] [riksprot] Wrong corpus document displayed.

[roger-mahler]

The following probems exist in riksprot notebook related to display of speech text:

• List of documents in find/browse has gender unknown for persons, but a gender exists in the displayed text.
• [Related to above]. All names in list of documents are the same (who id 0).
• The content of diisplayed text and associated topic(s) seem unrelated/off-topic.

• Further investigation reveal mismatch between speaker in document, and speaker in displayed text.

  Cause:

  • Analysis reveals that instead of fetching the nth speech the system fetches the nth utterance.
  • Also, when reading the member index, the who integer ID is lost, which causes item two above,

# FIXME [BUG] [riksprot] Wrong corpus document displayed.

https://github.com/welfare-state-analytics/welfare_state_analytics/blob/d39b3a93a5b92f646cefab77325f5d911c753aa9/westac/riksprot/parlaclarin/speech_text.py#L101