iislucas

Some conversations have multiple actions with no parent

1

In particular, there's about 1% of comments that have a section heading introduction at the same time as a comment addition (same revision & timestamp, but separated into 2 actions),...

Improve conversation edit/add distinctions

1

A few conversations that we seem to have mixed up one users edit with another's contribution, so the conversation looks like its one person talking to themselves, but actually it's...

Some comments initial text seems to be concat of a title and real text.

1

e.g. in conversation id: 232341942.8700.8700: The last comment starts: "Issue of the MonthThe great 'The/the debates' return. ". "Month" and "The" don't have any space. I think that's due to...

Add continuous integration testing using a .travis.yml file.

Current fails: * Search for phrases outside of individual article is too slow. e.g. last time someone said monkey. It returns the wrong results. Currently using SQL string search. *...

Support sending comments to the SuggestCommentScore API when moderation actions are taken. Distinguish between individual and batch actions. Make sure to include the reason for rejection.

To help avoid bugs and broken builds, lets setup [Travis integration](https://docs.travis-ci.com/).

Converter to tf.record improvements

1

* The current CSV to tf.record converter has hard-wired field selectors. These should be specified on the command line. * CSV is a bad format: there are multiple usually incompatible...

Create a token/embedding creation preprocessing pipeline using tf-transform

1

Issue: We currently depend on vocabularies, like glove embeddings, that are: 1. Weirdly biased (although when you backprop to the embeddings, their initial bias is not very relevant anymore), 2....

Examples: * Simple text-based word finder, and xor encoded problem (e.g. find "cat" or "dog" in string, but not both) * Training and metrics work.