Church code data cleansing process

[nurifreembd]

Discuss with Dale and Hilary steps needed to existing church codes, standardise the codes and auto generate codes for blank church codes.

[nurifreembd]

Church Code

CSV file names are currently constructed as follows,

ccceeettunique_number

where,

ccc = Chapman code for the county
eee = church code for the church concerned by the transcription
tt = transcription type, ie BA = baptisms
unique_number

FreeComETT gets this information as follows,

Chapman code is held against the county in the FreeREG backend database. So, knowing the county, the chapman code can be retrieved.

A new church code field was added to the church table in the FreeREG backend database a few weeks ago after request from the FreeComETT dev team. Prior to that the church code has never been stored anywhere and was a free choice of the transcriber creating the transcription.

FreeComETT picks up assignments created on the FreeREG backend and loads them to its own database to allow transcription of images stored in the FreeREG image server.

FreeComETT allows the creation of assignments within its own database without the assignment existing in the FreeREG backend. Rather than continuing to use a user generated church code, the team decided that it would a be a good idea to formalise and standardise the church code stored against a church in the backend database. Hence the addition of the church code field. This field was initialised for all churches with ‘198’. A totally illegal code. Church codes must be 3 characters long, no digits.

So far very few churches have church code assigned.

Currently when creating an assignment in FreeComETT, the user is asked to enter information required including the church code,

if the church code exists the system displays it to the user and allows change to it. FreeComETT enforces the 3 character rule but doesn’t test the code to see if it is unique
The backend church record is updated with the new church code.
The church code is stored in the FreeComETT database for the assignment and is used to create the CSV file name.

The first point above is controversial, since the code can be changed by the transcriber. It would be better to create church codes for all churches in the backend database

Some considerations,

church codes must be unique preferably across all churches but at least within a county.
They could be created automatically but,
there are existing CSV files which use a church code.
Could we not scrape these codes from existing CSV files?
Dale suggested he could create a list of all existing codes by chapman code and assign a code, something like,

Chapman code
Church
Used church codes
Church code to use to update the church table
STS
Hebron 1
HEB
HEB
STS
St Mary, Kingswinford
FIG
FIG
STS
St Mary, Winford
STW
STW
STS
St Therese, Kingswade
THG, THH, THF
THG

Some rules,
There can only be one church code for a given church
Church codes must be (at least) unique for a church in a given chapman code.
Church codes are 3 characters long.

Some consequences,
manually creating the of existing church code is likely to be a big task.
For churches with multiple existing codes, a unique code must be chosen.
Church codes within the chapman code must be tested unique
Once the code list has been created, it must be applied to the backend database. What information is required for this to work?
FreeComETT needs to be changed to use the church code in the church database and not allow change to it.
FreeComETT needs to be changed to remove the backend church table update to church code.
FreeREG backend needs to be changed to force entry of church code for a new/updated church table entry. Church code rules must be enforced.
WinREG users must be informed to use the new church codes.
FreeREG backend upload validation must be changed to verify church codes against chapman code/church to ensure use of correct church code
...