Skip to content

Conversation

@fscottfoti
Copy link
Contributor

Major file reorganization. This PR subsumes all the previous PRs. Also added tests to this - travis is likely to fail the first couple of times through - we shall see.

And all the alternatives, model, and all the new columns that are required to support it.  Only about halfway through at this point but hopefully gets a bit faster going forward.

Also added proper newlines to the csv file so they look ok outside of excel.

Also started keeping a todo list in markdown to keep track of anything that will need revisiting before the model is fully usable.
still need to polish one or two things before opening a PR though, but pretty close
for full time worker only at this point
only running for the full time worker person type at this time
there are a couple of issues - the window residual and the presence of X type not including chooser

other than that this is looking pretty good
I didn't realize all the segments in the non mandatory model had the same spec - this made things much easier (i.e. I didn't have to definite 250 variables for every segment).  Anyway this is now running and brings up a few questions which I will add to the readme.md.
a few caveats

1) this has no mode choice logsums

2) it uses fully random sampling

3) it's missing size terms

4) the escort trip type is not split into with and without child

...

but basic workflow is completely there - every tour now has a destination
not working yet as far as model specifications and such
need to ask a few questions about the time windowing
no new functionality here - just splitting up the one large file into 8-10 smaller files.  Now have a models directory and one file per model, as well as a tables directory and one file per table.  Remember this is only a convention, so mixing in other aspects of the simulation framework where appropriate is expected.
sharing some merges between models

also sharing the vectorized tour choice between mandatory and non_mandatory scheduling (should move to activitysim)
might take a couple of tries to get this working with travis
@fscottfoti
Copy link
Contributor Author

@jiffyclub Any idea why travis isn't running on this PR?

@jiffyclub
Copy link
Contributor

Looks like it's getting started now, dunno why it took a while.

@fscottfoti
Copy link
Contributor Author

So this is failing because there's a dependency on those dcm changes you made to UrbanSim. Can we update this repo to depend on master of urbansim or do you think we should do another urbansim release?

@jiffyclub
Copy link
Contributor

I don't want to release urbansim just yet because I need to figure out why the sanfran_urbansim test is failing. If you want to install urbansim from master on Travis take it out of the conda installation and add this line at the end of the installation section:

- pip install https://github.com/synthicity/urbansim/archive/master.zip

@fscottfoti
Copy link
Contributor Author

Sounds good - that should work.

Renaming variables and addding TODO markers where appropriate.  Also uncovered a bug in the simulation routine.  Will need to switch to tested activitysim code soon.
fscottfoti added a commit that referenced this pull request Mar 9, 2015
@fscottfoti fscottfoti merged commit 63fff26 into master Mar 9, 2015
@fscottfoti fscottfoti deleted the file-reorg branch March 9, 2015 17:52
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants