feat: Catch CSV parse errors and add validation notices

[aababilov]

If a CSV file failed to parse, there will be a clear user-visible
notice, e.g.:

        {
            "code": "csv_parsing_failed",
            "severity": "ERROR",
            "totalNotices": 1,
            "notices": [
                {
                    "filename": "stops.txt",
                    "charIndex": 278201,
                    "columnIndex": 2,
                    "lineIndex": 2148,
                    "message": "Length of parsed input (4097) exceeds the maximum number of characters defined in your parser settings (4096).",
                    "content": "..."
                }
            ]
         }

[aababilov]

Note that Univocity provides ExpandingCharAppender that can expand up to (Integer.MAX_VALUE - 8). We may switch to it in a separate PR if there is a use case for that (which I doubt because a CSV value longer than 4096 characters looks invalid to me).

[lionel-nj]

thanks for working on that @aababilov! Could you update RULES.md with this new notice description please?

Note that Univocity provides ExpandingCharAppender that can expand up to (Integer.MAX_VALUE - 8). We may switch to it in a separate PR if there is a use case for that (which I doubt because a CSV value longer than 4096 characters looks invalid to me).

@MobilityData/transit-specs could the GTFS specification add a requirement? "each cell value must be shorter than 4096 characters"? Unless there is a use case for as @aababilov suggested.

[lionel-nj]

Also @maximearmstrong you should take a look, see how his new notice code impacts the MobilityDatabase.

[aababilov]

thanks for working on that @aababilov! Could you update RULES.md with this new notice description please?

Done.

@MobilityData/transit-specs could the GTFS specification add a requirement? "each cell value must be shorter than 4096 characters"? Unless there is a use case for as @aababilov suggested.

Good question. This requires an extra PR.

[lionel-nj]

Thanks!

[timMillet]

@MobilityData/transit-specs could the GTFS specification add a requirement? "each cell value must be shorter than 4096 characters"?

Issue created to keep track https://github.com/MobilityData/gtfs-tasks/issues/118