# Structure


## Flow Chart
![title](_static/images/fwf-model-flowchart.png)

## File Directories

### bin
`/bluesky/fireweather/fwf/bin` is the operational directory, its where the code to run the model resides.
- all_fwf_run.sh 
- geo2topo_mv.sh

### scripts

`/bluesky/fireweather/fwf/scripts` is the operational directory, its where the code to run the model resides.
- run.py 
- geojson_maker.py 
- ds2json.py 
- index_generator.py
- build_intercomp.py
- wx_by_tzone.py
- wx_json.py
- wx_keys.py

### utils
`/bluesky/fireweather/fwf/utils`  is the model code  directory, all the scripts in ``master`` call on class methods and other functions in this directory.
- fwf.py
- read_wrfout.py
- geoutils.py

### data
`/bluesky/fireweather/fwf/data`   is the current forecast data directory, the directory has subfolders for each data type. 
- the key subfolder is ``/FWF-WAN00CG-01``  the current fwf forecast zarr files live there.
- the current fwf zarr forecast data is broken up into two groups ``hourly`` and ``daily`` the table below shows whats in each group.



### archive
``/bluesky/archive/fireweather/data/`` is the archive directory, it where copies of completed forecasts are stored as `.tgz`
- ``fwf-hourly-<domain>-YYYYMMDDHH.tgz`` for the hourly forecasts
- ``fwf-daily-<domain>-YYYYMMDDHH.tgz`` for the daily forecasts

### wrf
the model currently uses the 12-km and 4-km WRF 00Z but is easily adaptable to other domains and initialization times. 
- `/bluesky/working/wrf2arl/WAN00CG-01/YYMMDD00/` is the WRF directory where the model pulls in `.nc` files
- If you change to a new model domain you'll first need to run `../utils/timezone.py` to generate a tzone_ds.zarr file


## Required packages
Conda is used to manage python, its recommend to use conda to make the model work. 

- ``fwf-env.yml`` lives in the parent folder of the fwf repo
- to make a conda environment with all the required packages run the following code block

```
conda env create -f fwf-env.yml
```