JannisNe / timewise / 18199811002

87%

push

github

Refactor for v1-alpha (#151)

* move download config to separate file, make QueryType the only yhing needed, no QueryConfig anymore

* just define columns in queries

* move tables to separate module

* remove reduntant query instances

* remove reduntant query instances

* make query serializable by using a dtype map instead of types directly

* Use Chunk and Chunker

* make relative import

* add ampel

* make filesystem save data as csv

* add timewisealertloader

* timewisealertloader emits all rows per stock id

* rename to stock_id

* add comment

* add timewise alert supplier

* make ampel units discoverable

* rmove verbosity in query test

* add CLI for download

* fix public attribute

* return correct type

* fix generating tables

* correctly pass on dict

* adjust polling interval

* nicer hash

* make classvars

* use correct dtype from table

* add logging

* fix renaming columns

* use pandas to read cdv

* read stock id always as str

* copy of ZiDatapointShaper

* add TiDataPointShaper

* add TiDataPointShaper

* fix filename

* copy ZiMongoMuxer

* name must be literal

* add TiMongoMuxer

* use model field name dfault value

* do not combine dps for now

* add todo

* add copy of ZiCompilerOptions

* copy of ZTFIDMapper

* StockID must be int

* add timewise compiler options

* just use identity id mapper

* use pandas all the way

* fix column renaming

* ignore pandas settingwithcopywarning

* flatten file list

* check duplicate datapoints also for ra and dec

* ingest wise data structure, not per filter

* fix chunking

* add test ingest

* remove unsused statement

* add combine datapoints

* use hdf5 to store downloaded data

* add test for chunking

* make ampel config path relative

* use AMPEL_CONFIG_PATH

* add from_yaml to TimewiseConfig

* use backend to get files

* save data as fits, faster and smaller than csv

* adapt to not iterate over file

* better error

* fix saving to fits

* dont attempt column rename if no ... (continued)

1284 of 1467 new or added lines in 41 files covered. (87.53%)

1285 of 1468 relevant lines covered (87.53%)

0.88 hits per line