Four IMPS3 Sprints

[joereddington]

Piloting the next version of IMPS3 is taking a while so let's look into the list of features I wanted to build in the previous issue.

Sprint 1

• [x] Writers can press a button to get the pdf for their character.
• [x] index.php needs to have a good display for when it's not active (maybe put a picture in)
• [x] Bring in the pdf naming tools to the python. Break it down and make it testable.
• [x] Dashboard should respond to status - only allow pdfs if the right status is chosen

Sprint 2

• [x] Fix word count bug
• [x] Main page should be 'choose a character name or 'main'' so it's clear were people are meant to be.
• [x] Contents in the latex docs aren't appearing because you are deleting all the neccessary files. Should run twice
• [x] Books update automatically from the sprints list (python)
• [x] Etherpad tabs should open in new tab.
• [x] Generation of epubs when correct status is chosen

Sprint 3

• [x] Fix the epub publishing
• [x] Create seperate latex files for each slug to protect the pdfs.
• [x] write the document files to the latex sources directory NOT the python one
• [x] Fix the epub publishing
• [x] Strengthen the chapter tests
• [x] Fix the latex for multiple camps running at once (and think thought how many latex files we have)

Sprint 4

• [x] Germinate now works differently locally
• [x] Error messages for badly formed Seeds are now clear to user
• [x] index.php should complain if there isn't a match for the slug
• [x] Tests much faster (46 sec to 26 sec)
• [x] Bookdatabase now properly restores from files.

[joereddington]

16/05/22 06:26 to 07:04, WWW code +EQT

Deployed IMPS3 to server and, and part of this. Fixed some bugs.

• Read up on some git practices
• Obviously

I should be using a development branch for exactly this reason. I should also update the php on the AWS server. I'll wait until a week without camps.

Updating IMPS3 but can't work out why the php isn't doing the thing.

[joereddington]

09/06/22 12:42 to 15:08, Speeding up tests +EQT

Tests for IMPS3 are taking around 46 seconds.

I added setup and tear down code to make things a little easier to see. These are the worst offenders.

2.480 : test_etherpad.TestEverpad.test_download_all_books_that_start_with_slug 4.597 : test_chapter.TestChapter.test_writeoutputfilesWithCharacter 4.653 : test_chapter.TestChapter.test_populate_book_from_slug 4.666 : test_chapter.TestChapter.test_get_latex 4.666 : test_etherpad.TestEverpad.test_get_ordering 4.699 : test_bookdatabase.TestBookdatabase.test_cast_object_coming_out_of_database 4.704 : test_chapter.TestChapter.test_regress_latex 5.072 : test_bookdatabase.TestBookdatabase.test_add_book_to_database 5.226 : test_chapter.TestChapter.test_writeoutputfilesWithUnknownCharacter ~

It appears that at least one major offender is the singleton book code. It takes 5 seconds to get the book. I'd like it to work better by which I mean: I need it to work from the local database.

Check out these command lines:

git add --patch test_chapter.py

git stash pop git stash push -m "making singletonbook faster" git status

Next action is to examine the test for casting the file properly.

There's been a lot of refactoring and I'm reasonably happy that the code is in better shape.

[joereddington]

Some quick notes from being on a camp today

• IMPS3 bugs:
  • Case sensitive names
  • Germinate should allow particular overwrites.
  • Fix spellcheck again
  • Files without an owner.
  • Ability to include multiple characters (as in, someone might have steve and william)