ESTC raw csv prepicker

See change log

This is a package of two scripts for filtering the raw parsed ESTC csv.

OBS! The scripts were developed and tested with Python 3.6. Who knows what might happen with another version! Definitely won't be pretty with Python 2.x. No additional libraries needed.

ESTC raw csv precleaner allows filtering erroneous and duplicated data to the raw csv and outputting filtered data in the save csv format.

ESTC csv fieldpicker allows subsetting the csv and picking selected fields for further analysis. It maintains the same csv format, but only keeps selected fields (with value) and associated Cu-Rives numbers.

Pubdata cleanup starting data creates a csv in a different format that plugs into the start of the publisher data cleanup script. That one employs a custom method in ESTCMARCEntry -class in estc_marc.py and serves as an example of creating a starting data subset for a cleanup script. The MARC format for the different fields varies, so creating an universal function for this task doe not seem feasible.

Running the scripts

1. Run precleaner to remove duplicates, etc from the raw data.
2. Run fieldpicker to select fields of interest for further analysis down the ESTC pipeline.

ESTC raw csv precleaner prefilter_main.py

The purpose of this script is to allow filtering the raw ESTC csv. It removes various test records and and other garbage left in the csv. It makes sure that there are no duplicated Cu-Rives ids, and no entries that are missing and id. All the filtered entries are output to separate csv files to allow manual inpection.

Input

• Raw ESTC .csv -file produced by COMHIS/estc-xml2csv from input data at COMHIS/estc-data-originals.

Output

Output is csv with rows separated by tabs.

The script produces 3 output files (with file names configured in ./prefilter_conf.py):

• "Sane" ESTC entries. This is the one you want for starting point of cleaning scripts. Nothing else needed.
• Erroneous entries filtered out.
• Duplicated entries also filtered out.

There's also a separate table only containing the record_seq and estc_id mappings and sanity status.

Running the script

1. Set input and output file locations in prefilter_conf.py (symbolic links can help to avoid user specific paths)
2. Run prefilter_main.py (e.g. python3 prefilter_main.py); redone Apr 2, 2022 to ensure the use of the latest 2020 ESTC data dump; see prefilter_conf.py for the file paths
3. Use copy.sh to Copy the files from out/ into github repository: COMHIS/estc-data-verified/estc-csv-raw-filtered/; remember to push the updates

ESTC csv fieldpicker: fieldpicker.py

An ESTC fieldpicker that even your grandma could use. Allows you to filter the big ESTC csv and only keep the fields that are needed for your processing pipeline.

Input

ESTC csv in the same format as for the above script. Preferably one that has already been filtered for duplicates, etc. In practically all cases this should be estc_raw_sane.csv found here: COMHIS/estc-data-verified/estc-csv-raw-filtered/.

Output

Output is in the same format as the input, with only the user specified fields remaining.

Running the script

1. Create a config json file in ./cfg/ subdirectory. See for example pub260.json.
2. Run fieldpicker.py --conf [yourconf_file] (e.g. python fieldpicker.py --conf pub260.json)

Further utilities for specific fields

Pubdata cleanup starting data

The script is at create_pubdata_cleaning_starting_point.py. Configuration variables are set within the script file.

Input

Takes the output of either of the above scripts as input. Input should include all the information for MARC field 260.

Output

Creates a .csv table like this as output at the location set in the script file:

cu-rives	260_pub_loc	260_pub_statement	260_pub_time
(CU-RivES)N30954	Basil :	Printed by J.J. Tourneisen ;	MDCCLXXXIX. [1789]
(CU-RivES)N30954	Paris :	"Sold by Pissot, bookseller, Quai des Augustins,"	MDCCLXXXIX. [1789]
...	...	...	...