Skip to content

Commit af2db6a

Browse files
Ezekiel-KruglickMaxBenChrist
Ezekiel-Kruglick
authored and
MaxBenChrist
committed
Fixing TODOs (blue-yonder#305)
* add notebook with principal component analysis, closes blue-yonder#259 (blue-yonder#267) * add notebook with principal component analysis * change past and future to train and test * add benjamini-yekutieli procedure (blue-yonder#220) * add benjamini-yekutieli procedure * remove type conversion * remove space * add comparison with kolmogorov-smirnov * describe cells & add plt.show() * correct spelling mistake * refactored notebooks and removed constant features in BH (blue-yonder#298) * added y-1 baseline model to google stock notebook (blue-yonder#299) * added changes for version 0.10.0 * Make the .travis.yml more look like the example one from travis * Bugfix for preventing the subsequent spawning of Python processes, if… (blue-yonder#296) * Replaced depricated module sklearn.cross_validation by sklearn.model_selection * Simplify generation of DataFrame. * Automate statement concerning the number of extracted features. * The number of relevant classes should not be hard coded. * Removed another hard coded sample number * Bugfix for preventing the subsequent spawning of Python processes, if feature_extraction is called subsequently * Discard pool termination for single processing * Terminate multiprocessing pool. * use name of value column as kind (blue-yonder#304) * Travis test (blue-yonder#300) * Seperated out unit and integration tests * Run both test suites indepentendly * Use the build in cache by travis for pip * Try to use 2 cores * First install requirements, then the project * Try out the new stages build feature * Next try... * And anoter one... * Continue.. * ... * Caching the virtualenv did not help * renamed build stages * set n_jobs for pytest to auto * fix naming of test stages * use with notation for filehandling * set number of cpus to 2 * limit number of jobs to 2 by pytest argument * use write AND read mode for profiling file * add __init__.py to test folder * use tempfolder for profiling files * add changes for version 0.10.1 * Major improvements to candidate uniqueness function and testing thereof. Also pulled candidate generation out into separate method from find_motifs to allow separate testing. Took care of a couple todos
1 parent 0dd40ed commit af2db6a

35 files changed

+3198
-340
lines changed

‎.travis.yml

Lines changed: 49 additions & 30 deletions
Original file line numberDiff line numberDiff line change
@@ -1,35 +1,54 @@
1+
# Travis Build file for tsfresh
12
language: python
3+
# We do not use sudo here, so we can use docker containers, which have better caching
4+
sudo: false
5+
# We want the pip folder to be cached, to speed up installation
26
cache:
37
directories:
48
- $HOME/.cache/pip
5-
before_cache:
6-
- rm -f $HOME/.cache/pip/log/debug.log
7-
python:
8-
- '2.7.11'
9-
- '3.5.2'
109
install:
11-
- pip install --upgrade pip
12-
- pip install numpy
13-
- pip install scipy
14-
- pip install coveralls
15-
- pip install .
16-
- pip install -r requirements.txt
17-
- pip install -r test-requirements.txt
18-
- pip freeze
19-
env:
20-
- NUMBER_OF_CPUS=1
21-
script:
22-
# hack to set xdist to use only 1 node instead of 32
23-
- sed -i 's/\-n 4/\-n 1/g' setup.cfg
24-
- python setup.py test
25-
deploy:
26-
provider: pypi
27-
user: MaxChrist
28-
password:
29-
secure: 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
30-
on:
31-
tags: true
32-
distributions: sdist bdist_wheel
33-
repo: blue-yonder/tsfresh
34-
after_success:
35-
- coveralls
10+
- pip install --upgrade pip wheel setuptools
11+
- pip install numpy scipy coveralls
12+
- pip install -r requirements.txt -r test-requirements.txt
13+
- pip install -U .
14+
- pip freeze
15+
jobs:
16+
include:
17+
- stage: prepare cache
18+
script: true
19+
python: 2.7.11
20+
- stage: prepare cache
21+
script: true
22+
python: 3.5.2
23+
24+
- stage: Run unit tests
25+
script: py.test tests/units -n 2
26+
python: 2.7.11
27+
- stage: Run unit tests
28+
script: py.test tests/units -n 2
29+
python: 3.5.2
30+
31+
- stage: Run the full test suite
32+
script:
33+
- sed -i 's/\-n auto/\-n 2/g' setup.cfg
34+
- python setup.py test
35+
python: 2.7.11
36+
after_success:
37+
- coveralls
38+
- stage: Run the full test suite
39+
script:
40+
- sed -i 's/\-n auto/\-n 2/g' setup.cfg
41+
- python setup.py test
42+
python: 3.5.2
43+
44+
- stage: Deploy to pipy
45+
script: skip
46+
deploy:
47+
provider: pypi
48+
user: MaxChrist
49+
password:
50+
secure: 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
51+
distributions: "sdist bdist_wheel"
52+
on:
53+
tags: true
54+
repo: blue-yonder/tsfresh

‎CHANGES.rst

Lines changed: 19 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -4,6 +4,25 @@ Changelog
44

55
tsfresh uses `Semantic Versioning <http://semver.org/>`_
66

7+
Version 0.10.1
8+
==============
9+
- split test suite into unit and integration tests
10+
- fixed the following bugs
11+
- use name of value column as time series kind
12+
- prevent the spawning of subprocesses which lead to high memory consumption
13+
- fix deployment from travis to pypi
14+
15+
Version 0.10.0
16+
==============
17+
- new feature calculators:
18+
- partial autocorrelation
19+
- added list of calculated features to documentation
20+
- added two ipython notebooks to
21+
- illustrate PCA on features
22+
- illustrate the Benjamini Yekutieli procedure
23+
- fixed the following bugs
24+
- improperly quotation of dickey fuller settings
25+
726
Version 0.9.0
827
=============
928
- new feature calculators:

0 commit comments

Comments
 (0)