Welcome to Metrique’s documentation!¶
Contents:
Release Notes¶
0.3.0¶
- metriqued server dropped; metriquet -> github.com/kejbaly2/tornadohttp
- new MetriqueObject and MetriqueContainer Mapping classes
- refactored (moved into) MongoDBProxy and MongoDBContainer Mapping classes
- remaining metrique libraries merged into single metrique package
- using joblib for parallel job execution (not futures)
0.2.6¶
- auto-deploy and service management scripts added (metrique)
- mongodb replicaset support
- metriqued server failover pool support
- py26 no longer supported
- metriquet - generic, reusable tornado server class
- object consistent hashing bugfix
0.2.*¶
- metriquec.cubes module added with generic cube reference classes
- initial test framework added (pytest)
- plotrique imported into source repo
- user and cube profile api added
- authentication api added
- basesql activity import improvements (parallel support, single pass)
- object consistent hashing support
- cube refactors (use pandas.csv_import, etc)
- lots of docs added (still much more to go!)
- app universal default paths set (~/.metrique)
0.1.3¶
- server no longer runs cube extractions; instead, server exports save_objects and clients are expected to push data through the api
- refactored/updated client/server apis
- enhanced debuging
- basic auth
- SSL support
- tons of docstring updates
- update source code autodoc sphinx building
- csv now gets its URI on .extract() not __init__()
- basecube refactor
- built-in cubes: jkns, CSV, JSON, git
- client side celery periodic, automated extraction integration
- tons of client pandas/plotting additions to help with analysis of historical data, including a metrique result object (subclass of pandas.dataframe)
- enhanced logging
- new .distinct() method to return all unique values of a given field and given collection of documents
- json decoder for handing ObjectIds and datetimes
- JSONConfig now acts more like a dictionary
- version and requirements (pypi) info included
Autogenerated source code documentation: