Skip to content

Latest commit

 

History

History
123 lines (86 loc) · 6.88 KB

etpl_table_seed_guide.md

File metadata and controls

123 lines (86 loc) · 6.88 KB

Seeding the ETPL Table

This document contains steps for updating the data in the etpl database table. This requires a separate document because it requires multiple steps for standardizing the data, inserting it, and updating search token data.

Prerequisites

You will need the following:

  • Terminal
  • Python version 3.7 installed
  • pipenv package installed

Get CSVs and Clean Programs

  1. Download the latest programs and providers CSVs and name them as oldPrograms.csv and oldProviders.csv, respectively.
  2. In Terminal, in the d4ad root folder, run the commands below.

    Context: This formats the file names to the *_yyyymmdd.csv and moves them to the backend/data folder.

date=$(date '+%Y%m%d')
mv oldPrograms.csv "backend/data/programs_${date}.csv"
mv oldProviders.csv "backend/data/providers_${date}.csv"
  1. In the same d4ad folder, run the commands below. You may see a warning in the output, that's okay!

    Context: This creates the programs_yyyymmdd_merged.csv file, with DEGREEAWARDEDNAME, LICENSEAWARDEDNAME, and INDUSTRYCREDENTIALNAME columns added and SUBMITTERNAME and SUBMITTERTITLE removed (private data).

cd backend/data/program-credentials
pip3 install -r requirements.txt
python3 merge-credentials.py $date
  1. Open the programs_yyyymmdd_merged.csv file in your IDE. Delete any empty lines at the bottom of the file.

Combine Programs and Providers

  1. In the same program-credentials folder, run the commands below. If you are asked for a Postgres password, include the one you made for the user postgres when setting up the app. You may get an error that "the etplcombination table doesn't exist", that's okay!

    Context: This adds the _old suffix to the standardized CSV file to make way for the new one, and creates the combined_etpl_raw.csv file.

cd ..
mv standardized_etpl.csv standardized_etpl_old.csv
./combine-etpl.sh "programs_${date}_merged.csv" "providers_${date}.csv"

Standardize Combined Data

  1. In the same backend/data folder, run the commands below.

    Context: This will create the new standardized_etpl.csv and put it in the same folder.

./run-standardization.sh

6.1 To support the data migration to Credential Engine, we also need to create a modified form of standardized_etpl.csv that does not contain linefeed characters within the rows of the CSV file (This is a limitation in the import process to the system consumming this content for the data migration). To support this, we run an additional script to create standardized_etpl_for_data_migration.csv

python3 transform-for-migration.py

Create migrations to update Postgres DB

To learn more about database migrations and seed updates, see the README.

  1. In the same backend/data folder, run the commands below.

    Context: This uses the DBMigrate library to create the migration files (up and down SQL, JS), all prefixed with the current timestamp.

cd ../..
npm --prefix=backend run db-migrate create update-etpl -- --sql-file
  1. In the same d4ad folder, run the command below. Be sure to change the UP and DOWN file names to what was created in the previous step.

    Context: This will delete all rows from the current ETPL table and insert new rows from the latest CSV file, essentially overwriting it with new data. After running, you should see the UP and DOWN files now have thousands of lines of INSERT statements. It should also have the programtokens SQL statements added at the end of both files. This SQL will ensure that the tokens being searched on are up-to-date with etpl table changes. Please see decision_log.md #2020-08-12 for explanation of why we need this.

./backend/data/create_update_etpl_migrations.sh standardized_etpl_old.csv standardized_etpl.csv backend/migrations/sqls/INSERT-UP-FILENAME.sql backend/migrations/sqls/INSERT-DOWN-FILENAME.sql

  1. Create new up and down files with the same names but with the -TEST.sql suffix instead. 10.1. If your operation is just an update, add the following comment -- intentionally left blank to both files. 10.2. If your operation adds new columns, your DOWN file should have -- intentionally left blank but your UP file should have SQL code. Do this by first making a copy of standardized_etpl_test.csv and naming it standardized_etpl_old_test.csv. Then, add your new column(s) to the standardized_etpl_test.csv file in Excel and fill with dummy data. Finally, run the same script in Step 9 above with the test CSV and SQL files.

  2. Modify the migration .js file to conditionally load the -TEST.sql up and down files. Follow the instructions in the README section starting with "edit the corresponding .js file..." to see what code to replace.

  3. Delete the files that you will not need to push to Github. Run the following command in the backend/data folder.

rm standardized_etpl_old.csv && rm "programs_${date}.csv" && rm "providers_${date}.csv" && rm "programs_${date}_merged.csv"

Update the programtokens DB

  1. If your operation adds new columns, also add this to the TEST up-file.
delete from programtokens;

insert into programtokens(programid, tokens)
select etpl.programid,
       to_tsvector(coalesce(etpl.standardized_name, etpl.officialname)) ||
       to_tsvector(coalesce(etpl.standardized_description, etpl.description, '')) ||
       to_tsvector(coalesce((string_agg(soccipcrosswalk.soc2018title, ' ')), ''))
from etpl
         left outer join soccipcrosswalk
                         on etpl.cipcode = soccipcrosswalk.cipcode
group by etpl.programid;

Test the Changes

  1. Finally, run the automated tests in the d4ad folder with the command below.
./scripts/test-all.sh && ./scripts/build.sh && ./scripts/feature-tests.sh

Troubleshooting 1: If you get the error about a Postgres password not being of the right type, you may need to change your Postgres user permissions. This involves finding the pg_hba.conf file on your computer, opening it in your Terminal via Vim editor, and changing all the permissions to trust instead of sha_....

Troubleshooting 2: If Cypress feature tests fail, you likely need to just change some of the data that we are looking for in the Training Explorer. Check the video for what data is showing in these instances, and update the test to check for that same data. You'll see examples of this in past commits to update data.

  1. Add, commit, and push the requisite files (up and down SQL files, the up and down TEST SQL files, the updated JS migration file, and the updated CSV file). The continuous deployment will automatically run the script that finds this new migration and executes it, thus updating the Postgres database with the commands outlined in the migrations.