originial flash

Author: Ananya Joshi

_delphi_utils_python/delphi_utils/flash_eval/README.md

@@ -21,47 +21,8 @@ Types of outliers/changes FlaSH intends to catch are:
 - Changes in health condition [ex: new variant]
 
 ## Running FlaSH-eval
-In params.json, there are two parameters: flagging_meta which is a dictionary and flagging, which is a list of dictionaries. 
-Two key parameters in flagging_meta are:
-1. **remote**: True/False
-Are you running this system so that it's checking against the s3 filesystem or a local one? 
 
-True = S3, False = Local 
-
-Related Parameters: output_dir, input_dir
-
-2. **flagger_type**: flagger_df/flagger_io
-
-File regeneration is time-consuming. Should you regenerate all files specified or only those that need to be updated. For example, between different runs where only AR parameters are changed, the reference files can stay the same. 
-
-flagger_df: Regenerate All, flagger_io: Regenerate only necessary files 
-
-
-There are a few different ways to run the flagger. 
-
-1. What input data are you using? 
-Types: "api", "raw", "ratio" 
-- API: In params.json, flagging, set "sig_type" to "api", and the dataframe will be generated.
-- Raw/Ratio: Create a file, flag_data.py, in delphi_* for the indicator of interest that handles different sig_types as expected. See changehc/delphi_changehc/flag_data.py
-- Existing csv: Point to relevant location in 'raw_df' location
-
-- The dataframe should be as follows: 
-
-**Columns**: State Abbreviations [ak, ny, tx ...] & Lag Type. Total of 51 columns
-
-**Index**: Dates 
-
-So a sample dataframe would look like this:
-
-| -          | ak  | ny  | tx  | lags |
-|------------|-----|-----|-----|------|
-| 2021-12-03 | 100 | 123 | 45  | 1    |
-| 2021-12-04 | 30  | 20  | 78  | 1    |
-| 2021-12-03 | 300 | 323 | 90  | 2    |
-| 2021-12-04 | 90  | 40  | 100 | 2    |
-
-
-To run the flagging system, follow similar instructions as the validator readme copied below:
+First, run the indicator so that there are files for FlaSH to check. 
 
 You can excecute the Python module contained in this
 directory from the main directory of the indicator of interest.
@@ -85,7 +46,6 @@ the flagging system , as follows:
 
 ```
 env/bin/python -m delphi_INDICATORNAME
-env/bin/python create_df_process.py #this is up to you!
 env/bin/python -m delphi_utils.flagging
 ```
 
@@ -97,38 +57,12 @@ deactivate
 rm -r env
 ```
 
-You have a lot of flexibility for new functionality of the flagging module. 
-
 ### Customization
 
-All of the user-changable parameters are stored in the `flagging` field of the indicator's `params.json` file. If `params.json` does not already include a `flagging` field, please copy that provided in this module's `params.json.template`.
+All of the user-changable parameters are stored in the `flash` field of the indicator's `params.json` file. If `params.json` does not already include a `flash` field, please copy that provided in this module's `params.json.template`.
 
 Please update the follow settings:
-- flagging_meta
-  - "generate_dates": determines dates for parameters in flagging (below) are recreated daily
-  - "aws_access_key_id": for remote options,
-  - "aws_secret_access_key": for remote options,
-  - "n_train": the number of days used for training
-  - "ar_lags": the number of days used for the lag
-  - "ar_type": what type of autoregressive model do you want to use [TODO]
-  - "output_dir": location where files will be saved if using local filesystem 
-  - "flagger_type": flagger_df to regenerate all files or flagger_io to regenerate just the missing files
-- flagging: a list of dictionaries each with some of these params
-  - "df_start_date": start date of dataframe (used to create input df)
-  - "df_end_date": end date of dataframe (used to create input df)
-  - "resid_start_date": used to create the residual distribution 
-  - "resid_end_date": used to create the residual distribution 
-  - "eval_start_date": date range to create flags 
-  - "eval_end_date": date range to create flags 
-  - "sig_str": usually the signal name, used to create/save files
-  - "sig_fold": the name of the data source for organizational purposes
-  - "sig_type": the type of signal (raw, api, ratio) for organizational purposes
-  - "remote": are you using the local or S3 filesystem 
-  - "lags": how many lags do you want to consider. Consider if your signal does have lags and the role of backfill per signal 
-  - "raw_df": the location of the input dataframe
-  - "input_dir": location of relevant files to create the raw df
-
-
+- signals: a list of which signals for that indicator go through FlaSH. 
 
 ## Testing the code
 
@@ -163,12 +97,6 @@ or
 The output will show the number of unit tests that passed and failed, along with the percentage of code covered by the tests. None of the tests should fail and the code lines that are not covered by unit tests should be small and should not include critical sub-routines.
 
 
-## Code tour
-* run.py: sends params.json fields to and runs the validation process
-* generate_reference.py: generates the reference files related to a specific run 
-* generate_ar.py: generates the ar files related to a specific run 
-* flag_io.py: various functions to figure out which files need to be generated with specific parameters.
-* flag_data.py (local): generates the input dataframe (see application in runner.py)
 ## Adding checks
 
 To add a new validation check. Each check should append a descriptive error message to the `raised` attribute if triggered. All checks should allow the user to override exception raising for a specific file using the `suppressed_errors` setting in `params.json`.

_delphi_utils_python/delphi_utils/flash_eval/constants.py

@@ -1,5 +1,4 @@
-from datetime import date
-import pandas as pd
+"""Constants used in FlaSH."""
 
 #Regions considered under states
 STATES = ['ak', 'al', 'ar', 'as', 'az', 'ca', 'co', 'ct', 'dc', 'de', 'fl', 'ga',
@@ -9,10 +8,5 @@
             'or', 'pa', 'pr', 'ri', 'sc', 'sd', 'tn', 'tx', 'ut', 'va', 'vi', 'vt',
             'wa', 'wi', 'wv', 'wy', 'us']
 
-HOLIDAYS = pd.to_datetime(["1/1/2020","1/20/2020","2/17/2020","5/25/2020","7/3/2020","9/7/2020","10/12/2020","11/11/2020","11/26/2020","12/25/2020",
-"1/1/2021","1/18/2021","2/15/2021","5/31/2021","6/18/2021","7/05/2021","9/06/2021","10/11/2021","11/11/2021","11/25/2021","12/24/2021","12/31/2021",
-"1/17/2022","2/21/2022","5/30/2022","6/20/2022","7/04/2022","9/05/2022","10/10/2022","11/11/2022","11/24/2022","12/26/2022"])
-
-
 #HTML Link for the visualization tool alerts
-HTML_LINK = "<https://ananya-joshi-visapp-vis-523f3g.streamlitapp.com/?params="
+HTML_LINK = "<https://ananya-joshi-visapp-vis-523f3g.streamlitapp.com/?params="