Path4gmns

v0.9.10 comes with bug fix, new functionality, and performance improvement.

1. Fix bug on path cost unit for find_shortest_path() reported in Issue 58.
2. Find the shortest path per travel distance or travel time.
3. Introduce get_shortest_path_tree() to return the shortest path tree from a given source node.
4. Optimize the UE module by precluding zones with no valid demand, moving the check on each centroid into SPNetwork setup, and computing the relative gap in the postprocessing.
5. Offer users more flexibility to control UE convergency with the relative gap tolerance (i.e., the target relative gap), which is inspired by 57.

v0.9.9.post1 serves as a hotfix over v0.9.9. It reestablishes zone id cross-validation (between node.csv and demand.csv) and resolves network inconsistency by invoking find_shortest_path() and find_ue() in the same code snippet 51.

v0.9.9 comes with bug fixes, new functionality, and performance improvement.

1. Fix bugs in loading synthetic zones and synthesizing demand.
2. Separate demand loading from read_network().
3. Introduce a new public API read_demand() to unify demand and zone synthesis, and demand (and synthetic zone) loading.
4. Deprecate confusing perform_column_generation() and replace it with find_ue().
5. Introduce ODME.
6. Add support to load link toll from link.csv.
7. Optimize how to check if a key exists.

v0.9.8 fixes issues regarding accessibility output and supports recurring calls of run_DTALite(). Besides, it introduces **unit testing** and moves updates and implementation notes from README to docs.

v0.9.7 is a hotfix over v0.9.5 and v0.9.5 on the following issues.

1. Calling DTALite fails.
2. Synthesizing zones and OD demand is not functioning 41.

v0.9.6 brings several enhancements over its predecessor.

1. Eliminate the ultra-low-volume columns from the UE assignment, which leads to improved assignment results and much cleaner agent.csv.
2. Calculate the final UE convergency after the postprocessing, which provides the accurate assessment of each assignment.
3. Embed and support the latest [DTALite](https://github.com/asu-trans-ai-lab/DTALite), which includes all-new Logbook, enhanced scenario handling, improved I/O functionality, and so on compared to the existing [classic version](https://github.com/jdlph/DTALite).