Create JSON data files from Meteocode and Bulletin files

Programs in `src` directory

buildJSON.py : main program (TODO: allow command line calling)
buildRegionCodes.py : create data/codeRegions.json from data/regions.ont.txt and data/regions.que.txt
checkMeteocodeBulletin.py : sanity check between names of files in Meteocode and Bulletin directories
makeBulletin.py : create a bulletin using the jsRealB realiser (work in progress)
make_complete_json.py : create json files for all Meteocode and Bulletin file
parseMeteocode.py : parse a Meteocode file to produce a JSON structure
ppJson.py : pretty print of a JSON in a form that is more compact and readable than the standard Python pprint
stats.py : compute statistics on the fields of the json files and the vocabulary used in the bulletins

make_code_regions.py to build region mappings.
make_complete_json.py to parse bulletins and create JSON files for each bulletin
merge_all_jsons.py to create single JSONL file
compact_merged_json.py
1. trim to optimize format
2. partition to partition in train/dev/test
3. tokenize to tokenize texts in a given JSONL file

NB: for the moment, only a small directory was created for testing and developing

The variable baseDir should be in front of those directories
* stands for FP(TO|CN)MM.dd.hhmmZ
where