Rjf/gtfs final fixes

[robfitzgerald]

this PR contains the final set of fixes applied while working toward a successful run of BAMBAM for GTFS-Flex access modeling, which includes the change set from #158 which was cancelled after discovering it was missing a required fix present in this PR.

GTFS-Flex run fixes

• make negate key on destination filter optional for simplicity
• DestinationFilter comes from both config and query for mode-specific filtering
• fix bounds checking errors in multimodal_state_ops
• improved logging + error reporting during OMF import
• remove now-unnecessary allocation when calling state_model.contains_key and provide error context in multimodal_traversal_ops

simplify OMF Places mapping

these mappings can be quite long - there are over 2000 OvertureMaps categories total - and so expressing that directly within a bambam config file is noisy. this PR allows the user to alternatively define a CSV source for category mapping:

[plugin.output_plugins.model.models.opportunity_source.places_activity_mapping]
places_mapping_input_file = "mapping.csv"
places_category_column = "omf_label"
mep_category_column = "mep_labels"
mep_category_separator = "-"  # sep to use when more than one MEP label is produced

testing

1. unpack and copy these files into the base of the bambam repo:

omf-csv-test-assets.zip

overture-to-mep.csv
test_denver_omf.toml

2. run bambam:

% cargo build -r --manifest-path rust/Cargo.toml
% RUST_LOG=info ./rust/target/release/bambam -c test_denver_omf.toml -q query/denver_extent.json

3. a JSON and CSV output should appear with no errors logged to console or into the file contents, and the opportunities counted for each grid cell/mode/timebin come from OvertureMaps categories mapped to MEP categories as defined in teh overture-to-mep.csv file instance.

[robfitzgerald]

@admrtin this PR is ready for review. as in #158 i think my main goal for a review is for you to run the test described above. there's a note describing additional fixes. i haven't gone so far as to create new tests of the code changes and i'm ok with that choice for the most part, except for maybe the code that had the out-of-index error fixes. let me know what you think.

[admrtin]

The tests for the PR passed with no errors and I received the test-denver-result.csv and test-denver-result.json files. I'm good with merging this, and if we feel the need to add unit tests for the out-of-index error fixing later, we can.