ToDo Before Release

[krischer]

• [x] 5 - 10 MB datbenbank für tests, all modes (fwd, bwd, displ_only, fullfields)
• [x] rename
• [x] test coverage
• [x] fullfields modus
• [x] lanczos resampling in obspy
• [x] documentation, examples
• [x] parallel IO
• [x] lokaler webservice
• [x] refactoring
• [x] installation with setup.py
• [x] travis integration
• [x] PEP8 checks in tests
• [x] Rebase history to exclude large binary files
• [x] Test if only one file is present
• [x] Test without read-on-deman
• [x] Test for fullfield mode and test data
• [x] tests for strain only mode
• [x] force source tests
• [x] Finite with source time function test + srf reader
• [x] CMTSOLUTION reader, source from strike dip rake
• [x] Test _get_band_code()
• [x] Doctest for DB testing
• [x] Resampling test
• [x] Buffer test
• [x] Raise warning if depth of source/receiver is set in different modes
• [x] Format support for sources (QuakeML, NDK (once obspy has a new release))
• [x] either make the sem_derivative tests work or scrap them

[martinvandriel]

need to have some checking for source / receiver depth, because either of them depends on the db: receiver depth for bwd mode, source depth for fwd mode

[martinvandriel]

for fullfields mode: need to implement in axisem a version that saves displacement only for fwd, but on selected gll points and strains only for bwd. atm, fullfields stores both in both cases.

[martinvandriel]

fullfields mode always stores global wavefields, making the database 50MB big, so probably we don't test it, as strain only is more interesting for instaseis anyway