Week 2
Aug 7th 2023
Completing tasks from Friday discussion on pipe-breaks
add notes on msi and hp env in project.docs.com
Document dependencies adding in requirements.txt
Vannary's email
format of models in 2 formats (.sav and .pkl) - why?
(Van) I initially saved the model with .sav, as this is what I learned from my class. Deven, later on, changed it to .pklRequirements on compute-canada - pip freeze
(Van) For pipe-breaks, we used the 3.7 Python version. Please see the attached text file for lists of libraries used to train the model.
1. requirements.txt is a file I used to do pip install -r requirements.txt
2. list_of_packages.txt is a file containing all the packages installed in the environment that I used to train the pipe break models.
check the format of dates in tables
- The input & output format of csv in pipe-breaks seems to be year/month/day concatenated.
Aug 8th 2023
- Find diff btw xgbse_gqc and xgbse_gqc_2
- Things done on MSI:
- Created
pipe-breaks-transfer-learningand installed all requirements from transfer learning repo - Changes form HP transfer learning were pushed and pulled into MSI.
- Streamlit works with models generated by script_runner.py in
pipe-breaks.
- Created
- Pushed documentation on pipe-breaks to projects docs site.
- Debugging XGBSE algorithm on
pipe-breaksrepository. - Updating Docs
- Document that we should not edit auto-generated docs or scripts.
- Document about word2md.com
Aug 9th 2023
Working on documentation for workflow and deploying comment extractor.
Debugging XGBSE algorithm on
pipe-breaksrepository.Fix below warning while running xgbse_gqc_2 on
pipe-breaksrepository.C:\git\gqc\ai\pipe_breaks\common\processing.py:592: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
Aug 10th 2023
- Jake renamed old
gqc-utility-pythonrepo togqc-utility-notebooksand created new repogqc-utility-python. - All the docs sites have been updated with the latest links.
gqc-utility-pythonhas comment_extractor.py in it.-
gqc-utility-pythonneeds to have docs using Sphinx
-
- pipe-breaks
- Change pipe-breaks data from .csv to SQLite (later)
Aug 11th 2023
- Understand why histograms are displayed in a particular way after running
compare-results.py- Because of the bin interval being large. Here is the same plot with different bin intervals.

- Because of the bin interval being large. Here is the same plot with different bin intervals.
- Revert inputs / outputs file changes before pushing code changes to git for
pipe-breaks - Generalize hardcoded text like DNV, XGBSE in
pipe-breaks > training and predictionwith {utility} and {algorithm} - Make corresponding code changes to
loe-coe-appand resolve the errors - Sync environments in pipe-breaks, loe-coe-app and pipe-breaks-transfer-learning.
- (low) Look into Streamlit display of plots and graphs, such that they are user-interaction friendly like React graphs.