Skip to main content
Python Software Foundation 20th Year Anniversary Fundraiser  Donate today!

Python wrapper around tool ltlcross from Spot library

Project description

ltlcross wrapper

Python wrapper around the amazing tool ltlcross from Spot library. The tool compares LTL to automata translators.


The following libraries are needed for bokeh scatter plots (can be used in Jupyter)

  • bokeh (installs automatically by pip)
  • colorcet (installs automatically by pip)
  • jupyter_bokeh for rendering the plots in JupyterLab


python3 -m pip install -U ltlcross_wrapper


ltlcross_wrapper offers 3 classes:

  • Modulizer and GoalModulizer split a big ltlcross task into smaller ones, execute the small tasks in parallel, and merge the intermediate results into one final .csv, .log, and _bogus.ltl files. Always use GoalModulizer if one of the compared tools is GOAL!
  • ResAnalyzer parses the results of ltlcross, and implements several functions to analyze and visualize the results, mainly in Jupyter notebooks.


We need to specify the tools and file(s) with formulas which ltlcross should use. The tools are given as a dict whose items are pairs (name, ltlcross_cmd) where ltlcross_cmd is COMMANDFMT from man ltlcross.

tools = {"LTL3HOA"    : "ltl3hoa -d -x -i -p 2 -f %f > %O",
         "SPOT"       : "ltl2tgba"

Typical usage of Modulizer follows.

m = Modulizer(tools, formulas.ltl)

The above command splits the file formulas.ltl into several files with 2 formulas each, and uses 4 processes to run ltlcross on these small files in parallel. The number of processes can be controlled by setting processes in both the constructor and the function run(). The number of formulas in each small file can be changed by setting chunk_size in the constructor.

If a previous computation was interrupted for some reason, consider using


instead of The function m.resume() will skip computing the small tasks for which it already finds an .csv file with the result.

You can delete the final results and intermediate files, and rerun the computation by


By default, all intermediate files will be created in directory and the merged files modular.csv, modular.log, modular_bogus.ltl will be created in the current directory. The base modular can be changed to new_name simply by setting name=new_name. Otherwise, each filename/dirname can be changed by setting out_res_file, out_log_file, out_bogus_file, and tmp_dir.

Temporary files in ltlcross commands

Some tools need to read the input from (or store the output in) a temporary file that is different from the one that ltlcross expects. This file name has to be specified in the command for ltlcross. As the task will be processed in parallel (unless requested otherwise with processes=1), we can have a lot of race conditions. For this purpose, each process created by the Modulizer class set its own value into an environment variable called $LCW_TMP. You can then specify tools with commands like

"tool1" : ltl2tgba %f > $ && tool1 $ ...

GOAL & GoalModulizer

Always use GoalModulizer instead of Modulizer if you need to run GOAL in parallel. GoalModulizer requires additional parameter goal_root to specify the path to root directory of GOAL (contains gc, goal, GoalModulizer uses a unique binary and a unique shadow folder for each process that runs the task. The path to the unique binary is stored in environment variable $LCW_GOAL_BIN which you can use in specification of tools to compare. The goal_root part of the path should be omitted in the ltlcross command specification. Specifying GOAL is as simple as

'my_goal_batch` : '$LCW_GOAL_BIN batch "{goals commands}"'

If you are using the batch mode of GOAL, you often need to specify the names of the temporary files in the GOAL command. The $LCW_TMP is not recognized as a variable inside these commands. Use `echo $LCW_TMP` which is a nested shell evaluation inside the GOAL command (enclosed in two `). See an example of complementation performed by GOAL and simplified by Spot.

tools = {
    "piterman": "ltl2tgba -B %f > $ && $LCW_GOAL_BIN batch '$temp = complement -m piterman `echo $`; save -c HOAF $temp `echo $LCW_TMP.out;' && autfilt --small --tgba $LCW_TMP.out > %O
    "SPOT": "ltl2tgba -B %f | autfilt --complement > %O"
m = ltlcross_wrapper.GoalModulizer(goal_root="PATH_TO_GOAL", tools=tools, formula_file="MY_FORMS.ltl")

TODO: Explain ltlcross options

Results' Analyzer

See the usage notebook. Currently, bokeh scatter plots do not render directly on github so you might consider to see the notebook on nbviewer

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for ltlcross-wrapper, version 0.7.3
Filename, size File type Python version Upload date Hashes
Filename, size ltlcross_wrapper-0.7.3-py3-none-any.whl (25.1 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size ltlcross_wrapper-0.7.3.tar.gz (24.8 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page