Please do not hesitate to contact me at nicg at wudingcheng dot com if you have any questions or concerns.

• qtpy
• Matplotlib
• Numpy
• Pandas
• statsmodels
• cvxpy
• cvxopt

Before installing this package, the above requirements should be installed. For windows user, cvxpy package and its dependencies can be download at [Unofficial Windows Binaries for Python Extension Packages].

The Anaconda is recommended to be installed. TSAnalyzer can support both Python2 and Python3. TSAnalyzer is planning to pnly support Python3.

• Download this repo, and python setup.py install on common line prompt.
• pip install tsanalyzer

After that, the package creates a command tsanalyzer in your path.

For detail information and help, please refer to DOC.

For more information and support, please contact me directly.

Email: scholar at wudingcheng.com

• Support another neu format, traditionally time of neu is calculated by mjd = round(365.25 * (yearfraction - 1970.0) + 40587.0 + 0.1) - 0.5; date=date2mjd(mjd), this also support (int(yearfraction), yearfraction - int(yearfraction)) * days_of_years), see date.py yearfraction2mjd, fyear2date for detail;
• add continuous time-series back from the v1.x, the continuous means no discontinuities in time series and outliers;
• Fix bug for read repeated station time-series;

• Fix bugs for Header Convert Tool;
• Fix bugs for TSAnalyzer.models.reader;
• Fix bugs for some UI layout;

• Update L1 regularization for detection offsets and trendchanges;
• Offsets detection and output format;
• test scripts;

GPL