OSEP 6: change how we handle multiple district offices

[jamesturk]

It is probably about time to get rid of the contact_details schema, which causes a lot of confusion, and instead move to an offices schema.

Problems:

• updating/merging offices is 70% of the complexity of the people admin tools, it makes automatic updates nearly impossible in states with multiple district offices
• API v3 represents this data as offices, but since it is merging multiple contact_details based on key, it does so in a way that loses data
• openstates.org sometimes groups phone & address together incorrectly since it doesn't know which go with which from the schema itself

Rough roadmap:

• [x] write an OSEP
• [x] update people schemas & tooling
• [x] update the data in openstates/people
• [x] change openstates.org to use offices
• [x] change API v2 to use offices
• [x] change API v3 to use offices
• [x] drop contact_details from database/etc.

[jamesturk]

27

[jamesturk]

import sys, glob
import yaml

def convert_file(filename):
    with open(filename) as f:
        data = yaml.safe_load(f)
    offices = []
    district_num = 1
    for cd in data.pop("contact_details", []):
        note = cd.pop("note")
        if note == "District Office":
            cd["classification"] = "district"
            cd["name"] = f"District Office #{district_num}"
            district_num += 1
        elif note == "Capitol Office":
            cd["classification"] = "capitol"
        elif note == "Primary Office":
            cd["classification"] = "primary"
        offices.append(cd)
    if offices:
        data["offices"] = offices
    with open(filename, "w") as f:
        yaml.safe_dump(data, f, sort_keys=False)

if __name__ == "__main__":
    for name in glob.glob("*/*/*/*.yml"):
        print(name)
        convert_file(name)