A tool for remembering bash history

Sometimes we organize our codes in many repos and hence many directories. Remembering the directories is a nightmare. So, I created this simple tool.

Which you may refer here dirhistory tool .

 

Getting number of trees in a XGBoost model

 We can dump the number of trees in the xgboost model by below method:

Load the model and save it as a JSON file. Of course this step is not needed if the model was saved as a JSON already.

Then simply extract the num_trees information from the JSON file.

Below code snippet we may use for the same:

1. import sys
2. import json
3. import xgboost as xgb
5. if len(sys.argv) < 2:
6. print(f'Usage: {sys.argv[0]} ')
7. exit(1)
9. loaded_model = xgb.Booster()
10. loaded_model.load_model(sys.argv[1])
11. loaded_model.save_model('/tmp/a_model.json')
12. with open('/tmp/a_model.json', 'r') as fp:
13. jsonrepr = json.load(fp)
14. print(jsonrepr['learner']['gradient_booster']['model']['gbtree_model_param']['num_trees'])

We can run it as shown below (after saving the code snippet to a file xgb_tree_count.py):
python xgb_tree_count.py model-file-path

Making unsearchable pdf a searchable one

Try this command: $ ocrmypdf unsearchable.pdf this_will_be_searchable.pdf Here is the link for this interesting tool https://github.com/ocrmypdf/OCRmyPDF