Marco Soldati 9cc8eb4b8fc Fix type in exec config 10 Mar 2019 Marco Soldati a48d2425a52 Add 3 features config 08 Mar 2019 Federico Benvenuto 63b6b332a3c fixed the probability computation for the HybridLasso algorithm 08 Mar 2019 Federico Benvenuto 59329c93b8e changed the probability computation for the HybridLasso algorithm 08 Mar 2019 Marco Soldati 1dd05d4dfd2 redo prediction for whole period. 08 Mar 2019 Federico Benvenuto 0de0e9f5baa config file for the combo-solution (on the fly training and prediction) 08 Mar 2019 Federico Benvenuto d287496ec03 M Merge branch 'master' of https://dev.flarecast.eu/stash/scm/pred/all_machine_learning 08 Mar 2019 Federico Benvenuto 91b2a4349d8 remove the 'trained model' json file (containing a possibly large pickle object) from the server when we do on-the-fly training and prediction 08 Mar 2019 Marco Soldati 80b03837244 use new 10 features configuration in test scripts 08 Mar 2019 Marco Soldati 9f998139716 Update 10 features configuration 08 Mar 2019 Marco Soldati b4ee3e0ac29 issuing cadence should take into account "3" and "03" values. 08 Mar 2019 Federico Benvenuto af5d1a7c836 With this modification the HybridLasso method produces a supplementary probability forecast. 07 Mar 2019 Federico Benvenuto c58758d6ec3 Corrected some minor issue and added comments. 07 Mar 2019 Federico Benvenuto c650e13fdf0 added the read_model method, which is used -in place of 'get_algo_model' - in the main file. 07 Mar 2019 Federico Benvenuto fb9601c01e9 This modification allows to the label data frame to be empty. This can be the case of realtime forecast. 07 Mar 2019 Federico Benvenuto 6f716a47d1b Changed the way 'fc_id' is added to the feature data frame. 07 Mar 2019 Federico Benvenuto 4f701cf70f0 fixed a bug: read 'properties' instead of 'active_features' in the trained model 07 Mar 2019 Federico Benvenuto 1370033716f Added offline keyword to deal with combo-solution. With this change we can (write and) read the model offline. 07 Mar 2019 Federico Benvenuto 13e872b1edc Fixed a bug in the date management. Removed the use of "additional features" (they do not necessarily exist in realtime data). Added a unique identifier to the name of the algorithm. 07 Mar 2019 Marco Soldati d10e7c12e27 Add 10 features config 07 Mar 2019 Federico Benvenuto 3784bae915c renamed params.son 07 Mar 2019 Federico Benvenuto bccef945e3c move params.son to test_config 07 Mar 2019 Federico Benvenuto eb377837557 This file serves as an example to run the combo solution: note that the when the phase is set to combo, additional features are set to false. 06 Mar 2019 Federico Benvenuto 77e8501ef91 added the combo solution to the main file. It is active when the config file contains 'combo' in the 'phase' field. It allows on the fly training and prediction. 06 Mar 2019 Federico Benvenuto be3a5ccbce9 mange the 'combo' phase keyword: within the engine it behaves as the training phase, but it is managed in a different way in the main.py allowing on-the-fly training and prediction 06 Mar 2019 Federico Benvenuto a06fac37267 added combo_solution.py to check available features and perform training and prediction on the fly 06 Mar 2019 Marco Soldati cdcc9911ef1 Fix broken LOG prefix. 06 Mar 2019 Marco Soldati 4a2b51b1fac cleanup new algorithm configuration 04 Mar 2019 Marco Soldati 41657c5611c Remove exit statement. 04 Mar 2019 Marco Soldati b1acabdad7f Add log statement 04 Mar 2019 Marco Soldati dc2344adbeb Add warning for future calls 30 Jan 2019 Marco Soldati 98ff3f2fdcb update test_all_main for development purposes only 30 Jan 2019 Marco Soldati b1d937e7fff Add config für ssa_algo 30 Jan 2019 Marco Soldati 85f0b0286c2 Get rid of warnings framework and replace by logger. some cleanup (together with Federico) 30 Jan 2019 Marco Soldati c898c3450c7 Try to fix yet another bug 14 Jan 2019 Marco Soldati 82656b133ac Try to fix yet another bug 14 Jan 2019 Marco Soldati 234327d43ef Try to fix yet another bug 14 Jan 2019 Marco Soldati c9d17c187f7 testing method for convenience 11 Jan 2019 Marco Soldati 342145ebcc9 Fix some issues with empty data frames 11 Jan 2019 Marco Soldati 46eaff0491e housekeeping 11 Jan 2019 Marco Soldati de29524d51d Prediction phase always took the wrong configuration. 03 Dec 2018 Marco Soldati a260fc5c7a8 Switch to experimental (faster) property reading 03 Dec 2018 Marco Soldati de6be1da5a8 Disable write prediction 03 Dec 2018 Marco Soldati 7e7cc5935a0 Prepare experimental properties extraction 03 Dec 2018 Marco Soldati 789bc53a596 Reset read_properties method, 03 Dec 2018 Marco Soldati 7b5bc2f2666 Make sure to create log dir 02 Dec 2018 Marco Soldati a3b0c1680ec Remove second call to property service. 02 Dec 2018 Marco Soldati 3fe1a5905d9 Completely refactor flarecast_engine.py in order to better control configuration. 01 Dec 2018 Marco Soldati 009957309f5 Add the full test runner 01 Dec 2018 Marco Soldati 169d9f64013 Add missing slash to /ui/ check 24 Nov 2018 
Marco Soldati
Federico Benvenuto