Luke Van Seters

Currently we expect arbitrages to look like Token 1 --> Pool A (100%) --> Token 2 --> Pool B (100%) --> Token 1 But it's possible to do something like...

Right now we use a CSV COPY approach for writing blocks and classified_traces This speeds up writing by ~25% We should implement the same for the rest of the tables

Due to an accidental comma, inputs are currently written to the DB as a 1 element list Need to remove the comma

Right now the protocol field for classified traces says `Protocol.unknown` etc When it's missing, it also writes the string None None should be a postgres null And the prefix shouldn't...

This transaction https://etherscan.io/tx/0x57e8b6fa4c7ae3d8fb2ce274a19dbc54579208ac91ee9be12162f6d4e3d9e9a1 includes a swap with the UniV3 SHOPX pool (`0xe4b82df044511a7162bea1f11119f6d1283a43a8`) at trace address 14 Action [68] of this transaction https://etherscan.io/vmtrace?txhash=0x57e8b6fa4c7ae3d8fb2ce274a19dbc54579208ac91ee9be12162f6d4e3d9e9a1&type=parity mev-inspect fails to decode this swap with the...

Add support for Slack emojis

3

Products like Slack use https://github.com/iamcal/emoji-data as the standard for alias-emoji pairs. This adds support for missing aliases to the emojis.json file.

Use `.to(device)` in place of `.cuda()` to support CPU and other devices

1

Currently only CUDA-enabled devices can run Dreamtalk. However, CUDA-specific logic is only used in a handful of places. This PR: - Adds an argument `--device` to specify a device to...

The constructor for `CVModel` in `TopicMembershipModel` is incorrect and is causing adatest to throw errors when generating topics: ``` TypeError: CVModel.__init__() missing 2 required positional arguments: 'embeddings' and 'labels' ```...