-
…new refitting logic, and revert back to original reference values for the unit tests
-
The "refit_if_failed" parameter instructs ctlike to perform a refit in case of a fit failure. Fit failures are either a stalled fit, an exhaustion of the number of iterations, or a significant difference between the number of observed and predicted events.
-
The Crab test event data are now derived from the Prod2 South_0.5h IRF with an offset angle of 0.5 deg and a simulation radius of 3 deg.
-
The outcome of the unit tests depends on the presence of the CTOOLS environment variable. If the CTOOLS environment variable is not present the unit test should fail, otherwise it sould succeed.
-
All likelihood ctools now derive from the ctlikelihood base class. This avoids code duplication and provides all likelihood tools with a uniform interface.
-
Note that this breaks compatibility with Python < 2.6.
-
This command was run: autopep8 --select E101,W191 --in-place --recursive . It fixes only these two issues: - W191 indentation contains tabs - E101 indentation contains mixed spaces and tabs
