phantop/quatalog-scraper
1
0
Fork 0
mirror of https://github.com/quatalog/quatalog.git synced 2024-11-08 08:04:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

add header to csv

Browse source
This commit is contained in:
powe97 2023-07-29 17:17:48 -04:00 committed by GitHub
parent 9b877abe66
commit 9ec4198155
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 2 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
.github/workflows/scraper.yml vendored
View file

@ -144,12 +144,13 @@ jobs:
- name: Create CSV file - name: Create CSV file
run: | run: |
echo '"Instructor","Course"' > quatalog-data/courses.csv
cd static-generated/courses cd static-generated/courses
grep -oe '<li>.*</li>' *.html | grep -oe '<li>.*</li>' *.html |
sort -u | sort -u |
grep -ve "-[0-9]9[4-7][0-9]" -e "-[0-9]9[89]0" -e "-[0-9]00[1-9]" -e "USA[RF]" -e "ADMN" -e "USNA" | grep -ve "-[0-9]9[4-7][0-9]" -e "-[0-9]9[89]0" -e "-[0-9]00[1-9]" -e "USA[RF]" -e "ADMN" -e "USNA" |
sed -e 's/^/"/' -e 's/.html:<li>/","/' -e 's#</li>#"#' -e 's/"\([^"]*\)","\([^"]*\)"/"\2","\1"/' | sed -e 's/^/"/' -e 's/.html:<li>/","/' -e 's#</li>#"#' -e 's/"\([^"]*\)","\([^"]*\)"/"\2","\1"/' |
sort > ../../quatalog-data/courses.csv sort >> ../../quatalog-data/courses.csv
- name: Push CSV file - name: Push CSV file
working-directory: quatalog-data working-directory: quatalog-data