how to get the dump data of stackoverflow and train it from scratch?

[SeekPoint]

No description provided.

[davidmezzetti]

Once codequestion is installed, the following steps will help build the model from scratch. I'll keep this issue open to also document this in the README:

1.) Download files from Stack Exchange: https://archive.org/details/stackexchange

2.) Place selected files into a directory structure like shown below (current process requires all these files).

stackexchange/ai/ai.stackexchange.com.7z
stackexchange/android/android.stackexchange.com.7z
stackexchange/apple/apple.stackexchange.com.7z
stackexchange/arduino/arduino.stackexchange.com.7z
stackexchange/askubuntu/askubuntu.com.7z
stackexchange/avp/avp.stackexchange.com.7z
stackexchange/codereview/codereview.stackexchange.com.7z
stackexchange/cs/cs.stackexchange.com.7z
stackexchange/datascience/datascience.stackexchange.com.7z
stackexchange/dba/dba.stackexchange.com.7z
stackexchange/devops/devops.stackexchange.com.7z
stackexchange/dsp/dsp.stackexchange.com.7z
stackexchange/raspberrypi/raspberrypi.stackexchange.com.7z
stackexchange/reverseengineering/reverseengineering.stackexchange.com.7z
stackexchange/scicomp/scicomp.stackexchange.com.7z
stackexchange/security/security.stackexchange.com.7z
stackexchange/serverfault/serverfault.com.7z
stackexchange/stackoverflow/stackoverflow.com-Posts.7z
stackexchange/stats/stats.stackexchange.com.7z
stackexchange/superuser/superuser.com.7z
stackexchange/unix/unix.stackexchange.com.7z
stackexchange/vi/vi.stackexchange.com.7z
stackexchange/wordpress/wordpress.stackexchange.com.7z

3.) Run the ETL process

python -m codequestion.etl.stackexchange.execute stackexchange

This will create the file ~/.codequestion/models/stackexchange/questions.db

4.) Build word vectors

Currently, the model is using BM25 + fastText for indexing, should this change this step isn't necessary

python -m codequestion.vectors

This will create the file ~/.codequestion/vectors/stackexchange-300d.magnitude

5.) Build index

python -m codequestion.index

After this step, the index is created and all necessary files are ready to query.

[SeekPoint]

it works!

nice job