federicoaureliano / aiutare Goto Github PK
View Code? Open in Web Editor NEWAutomated Analysis, Regression, and Evaluation
Automated Analysis, Regression, and Evaluation
Otherwise, all functions use the version of config as written to the bin.benching.config file during the PREVIOUS call to run
I've only done relational database design, so this will be new for me too:
Will need SAT and SMT subdirectories under:
python pathlib
Using MongoDB (https://api.mongodb.com/python/current/tutorial.html)
Should all be handled in each program's output2result function. The MongoEngine schema for each program should also be stored in this file.
Very sporadic... tricky to replicate.
The user should be able to specify which solvers to run from the command line; this is more efficient than the current method of editing bin/bench.py before every run.
One other option would be to have a separate JSON file (e.g. "solvers.json") containing all the command line input to be run. Will discuss.
Needs to be more flexible than stringfuzz (https://github.com/dblotsky/stringfuzz), which only generates SMT-LIB string problem instances.
To learn about fuzzing:
https://www.fuzzingbook.org/
First, move the SAT and SMT example files out of the tool's repository.
Then, create a more streamlined system for a user to feed prepare.sh a config file (possibly pointing to other necessary files) instead of moving files into the bin/categories dir.
Finally, calling run.sh and cleanup.sh ideally shouldn't need sudo access.
For each solver, we can use linear regression to generate a solver of the form “expected runtime = b0x + b1y + b2z + error”, where variables x, y, z correspond to certain extracted features such as #clauses, #variables, and average clause length.
Each model can also have a R^2 (coefficient of determination) attached which signifies how confident we are in the model's accuracy.
For complicated/nuanced features, such as "is any one variable in >50% of all clauses?", we can introduce binary variables into the linear regression model to further improve accuracy in all cases.
Automatically clone & build specified versions (indicated by commit hash) to run.
"git clone [repo URL]" + "git checkout [commit hash]"
(https://stackoverflow.com/questions/12256137/cloning-an-older-version-of-github-repo)
Using MongoDB (https://api.mongodb.com/python/current/tutorial.html) + Vega-Lite (https://vega.github.io/vega-lite/)
What kind of views would be useful?
The output2result functions will need to handle their own writing to the database, since we may be looking for different features from each solver/saving results will no longer be generic across all categories.
Run 2+ “control” solvers on each benchmark and calculate deviation based on their difference in runtime; somehow factor this into overall standard deviation and plot the line of significance on the performance graph.
Ideally these "real" problems have proven correct answers which we can store in the database and compare solvers against.
ERROR:
python3 aiutare.py
Process Process-1:
Traceback (most recent call last):
File "/usr/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/lib/python3.6/multiprocessing/process.py", line 93, in run
self._target(*self._args, **self._kwargs)
File "/home/fmora/Documents/software/finnbarr/categories/smt/schemas.py", line 46, in write_instances
if not Instance.objects(filename=instance):
File "/home/fmora/.local/lib/python3.6/site-packages/mongoengine/queryset/manager.py", line 37, in get
queryset = queryset_class(owner, owner._get_collection())
File "/home/fmora/.local/lib/python3.6/site-packages/mongoengine/document.py", line 208, in _get_collection
db.client.is_primary:
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/mongo_client.py", line 890, in is_primary
return self._server_property('is_writable')
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/mongo_client.py", line 722, in _server_property
writable_server_selector)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/topology.py", line 226, in select_server
address))
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/topology.py", line 184, in select_servers
selector, server_timeout, address)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/topology.py", line 200, in _select_servers_loop
self._error_message(selector))
pymongo.errors.ServerSelectionTimeoutError: No primary available for writes
Running 44 total commands
Process Process-2:
Traceback (most recent call last):
File "/usr/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/lib/python3.6/multiprocessing/process.py", line 93, in run
self._target(*self._args, **self._kwargs)
File "/home/fmora/Documents/software/finnbarr/bin/benching/run.py", line 34, in monitor_database
with db.watch([{'$match': {'operationType': 'insert'}}]) as stream:
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/database.py", line 530, in watch
batch_size, collation, start_at_operation_time, session
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/change_stream.py", line 85, in init
self._cursor = self._create_cursor()
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/change_stream.py", line 166, in _create_cursor
explicit_session=self._session is not None
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/change_stream.py", line 137, in _run_aggregation_cmd
client=self._database.client)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/pool.py", line 584, in command
user_fields=user_fields)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/network.py", line 158, in command
parse_write_concern_error=parse_write_concern_error)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/helpers.py", line 155, in _check_command_response
raise OperationFailure(msg % errmsg, code, response)
pymongo.errors.OperationFailure: Cannot use snapshots until replica set is finished initializing.
Traceback (most recent call last):
File "aiutare.py", line 59, in
main()
File "aiutare.py", line 52, in main
analyze()
File "/home/fmora/Documents/software/finnbarr/bin/analyze.py", line 134, in analyze
iterate_results(schemas, consensus_dict, counts_dict, times_dict)
File "/home/fmora/Documents/software/finnbarr/bin/analyze.py", line 48, in iterate_results
for result in schemas.Result.objects():
File "/home/fmora/.local/lib/python3.6/site-packages/mongoengine/queryset/manager.py", line 37, in get
queryset = queryset_class(owner, owner._get_collection())
File "/home/fmora/.local/lib/python3.6/site-packages/mongoengine/document.py", line 208, in _get_collection
db.client.is_primary:
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/mongo_client.py", line 890, in is_primary
return self._server_property('is_writable')
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/mongo_client.py", line 722, in _server_property
writable_server_selector)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/topology.py", line 226, in select_server
address))
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/topology.py", line 184, in select_servers
selector, server_timeout, address)
File "/home/fmora/.local/lib/python3.6/site-packages/pymongo/topology.py", line 200, in _select_servers_loop
self._error_message(selector))
pymongo.errors.ServerSelectionTimeoutError: No primary available for writes
One .sh file for solvers; give the user the option to install all of them or only a selection.
One .sh file for benchmarks; probably should warn the user of how large the zipped & unzipped files are (1.3 GB and 13.7 GB).
All contenders should be by Nick's recommendation ("Minisat, manySat, Glucose, MapleSat") or gold performers at SAT Comp 2017.
Ideally, the different "tracks" of SAT Comp will mean that certain solvers will behave very differently on different problems, giving us an interesting collection of results to analyze.
Need to read through SAT Comp 2017 proceedings to decide which solvers are valid & CDCL-based.
For example, a SAT problem has features such as:
which can be stored in the database and included in graphs on request.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.