Repository navigation
fix(demo): replay a captured Feodo Tracker poll - #420
Merged
Merged
Conversation
The demo pack's feodo-c2 fetcher had no captured poll, so verify.json exempted it as waiting and the fetchers check skipped it. The pack replays a captured poll like demo-theatre's adsb-tracks and ais-tracks replays. The fetchers check runs. - events/feodo-c2/demo.log is one poll (5 servers) of https://feodotracker.abuse.ch/downloads/ipblocklist.json. events/capture.py wrote it with the source's URL and headers. PROVENANCE.md records the licence. Section 3.4 of the abuse.ch terms puts Feodo Tracker under CC0 terms for commercial use, so no licence text is copied in. - feodo-c2.yaml has a replay unit pointing at demo-replay. replay_fetch.py serves it. The live blocklist unit still polls the tracker. The replay returns no servers even when the tracker has some. - replay_fetch.py reads an empty rows.at as the whole body, because the blocklist is a top-level JSON array. It accepts the demo's empty pack name. - verify.json drops the waiting block. The demo-feodo-c2 zero_ok entries stay until the demo run shows which tiles read non-zero. A dump unit has one step and no window (dfe-fetcher crates/rest/src/shape/window.rs), so the cold-start lookback filters nothing. The blocklist's dates are strings. The replay only moves integers within a day of the poll, so every value is served as captured. The replay runs in the published demo. make demo has no demo-replay service, so its feodo-c2 replay unit has nothing to poll.
The committed feodo-c2 source has just its live blocklist unit. make demo runs no demo-replay service, so it avoids polling an unreachable replay unit. replay_fetch.with_replay_unit builds the replay unit from the live unit. The demo-replay service and the published demo's source seeding use it. The verify.json zero_ok list drops its demo-feodo-c2 entries.
A replay unit without rows.at is rejected unless its decoder is json_array. That decoder reads the whole body as rows. Otherwise, a document unit would load and land twice. It would never be blanked when the feed has rows. docs/evaluating.md says feodo-c2 rows land in dfe.feodo-c2. A seeding test checks that a replay unit is added to each fetcher with captured polls. It checks that the others are unchanged.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The demo pack's feodo-c2 fetcher had no captured poll. verify.json exempted it as waiting, so the fetchers check skipped it. The pack replays a captured poll as demo-theatre does, and the fetchers check runs.
A dump unit has one step and no window (dfe-fetcher crates/rest/src/shape/window.rs), so the cold-start lookback filters nothing. The blocklist's dates are strings. Replay only moves integers within a day of the poll, so every value is served as captured.