- Nov 30, 2023
-
-
vlorentz authored
It's unnecessarily verbose and distracts from the actual error
- Nov 29, 2023
-
-
David Douard authored
-
David Douard authored
-
- Nov 27, 2023
- Nov 20, 2023
-
-
Jérémy Bobbio (Lunar) authored
When we had no results, we used to return a list with an empty string. We now skip blank lines entirely. Queries returning no results will therefore now return an empty list. Closes #4790
-
SimpleTraversalView.prepare_response sets: self.nodes_stream = self.rpc_client.Traverse(self.traversal_request) which sent a request then did not stream from it because it made its own request and streamed from that instead.
-
- Nov 15, 2023
-
-
-
Jérémy Bobbio (Lunar) authored
Use `RPCClient._get()` instead of `RPCClient.get()`.
-
Jérémy Bobbio (Lunar) authored
We used to allow users to create a RemoteGraphClient with any URL. This gets surprising results if the HTTP server is up, but the location is wrong. As the HTTP Graph API will return 404 for unknown SWHIDs, a client can fail to notice the configuration problem and silently get wrong results. Instead, we now query the `/stats` endpoint on instanciation and raise exceptions in case of 404, connection error or unexpected content in stats results. Closes #4789
-
- Oct 31, 2023
-
- Oct 23, 2023
-
-
vlorentz authored
Prevent timestamps in node properties from being shifted according to the timezone WriteNodeProperties is being run in. Due to our ORC exports using the `timestamp` instead of the `timestamp with timezone`, reader and writer need to agree out of bound on the timezone used in files they exchange. However, we don't do this: * `swh-dataset` uses pyorc, which uses the C++ ORC library, which assumes users (us) always write in GMT * `swh-graph` uses the Java ORC library, which assumes the system timezone (or `$TZ` if set) So when reading with a non-UTC timezone, the Java ORC library interprets timestamps in the dataset as being in the local timezone, and converts them to UNIX timestamps (number of seconds since epoch); then we use these converted timestamps and write them to `.property.author_timestamp.bin` and `.property.committer_timestamp.bin`. This commit regenerates the example graph to have the correct timestamps. It also applies the 39ed0d17 change that removes useless padding at the end of all property files. Resolves #4788
-
- Oct 16, 2023
-
-
Stefano Zacchiroli authored
Rationale: this way we showcase how to obtain the successors of a given node, not necessarily the current one, which might be needed when jumping across forward/backward graphs.
-
Stefano Zacchiroli authored
This is a variant of the preexisting ReadGraph.java, that was printing only the forward graph via SwhUnidirectionalGraph. ReadBiGraph uses SwhBidirectionalGraph instead and explicitly use forward/backward graph accessors before printing edges. ReadBiGraph also explicitly prints node IDs, to show that the node identifiers are consistent across both graph representations, allowing to jump back and fort between the two graphs.
-
- Aug 21, 2023
-
-
vlorentz authored
-
- Aug 14, 2023
-
-
vlorentz authored
-
- Aug 12, 2023
-
-
vlorentz authored
It crashes on the production dataset
-
- Aug 07, 2023
-
-
vlorentz authored
-
- Aug 03, 2023
- Aug 02, 2023
-
-
vlorentz authored
-
- Jul 20, 2023
-
-
Antoine Lambert authored
-
- Jul 13, 2023
-
-
vlorentz authored
-
- Jul 12, 2023
- Jul 11, 2023
-
-
vlorentz authored
-
- Jul 10, 2023
-
- Jul 07, 2023
-
-
Tommaso Fontana authored
added Node2Type, its tests, and an bin to convert the .node2swhid.bin file to the new .node2type.bin
-
David Douard authored
It now needs types-click which is indeed a dependency of swh.core[testing].
-
- Jun 28, 2023
-
-
Stefano Zacchiroli authored
as suggested by Nicolas Dandrimont and Tommaso Fontana in !324
-
Stefano Zacchiroli authored
This serves two purposes: 1) Enable to compare visit results between the implementation currently deployed and webgraph-rs. FWIW, I verified this for the snapshow SWHID in the example and indeed it returns the very same 344 SWHIDs returned by the swh-graph Web API for that SWHID (although in a different order, which is OK). 2) Showcase how to use the Node2SWHID map which was previously not used in the example.
-
Stefano Zacchiroli authored
In particular we now also log: - the actual files from which stuff is loaded from (useful in case of errors from webgraph-rs, as they are quite terse at the moment) - the values lookedup/resolved, e.g., nodes and SWHIDs, for better interactive inspection of what is happening
-
Tommaso Fontana authored
-
- Jun 27, 2023
-
-
-
Stefano Zacchiroli authored
-