annotate.sh

#!/bin/bash

if [[ -z "${OPEN_AI_API_KEY}" ]]; then
  echo "The openai api key env var is not set, please set a value and reinvoke"
  exit 1
fi

# do google authentication to go through the project setup for gemini
gcloud auth application-default login


# login in huggingface hub to get codebert and save token on machine
huggingface-cli login

echo "Enter your programming language, e.g. C, go, java, js, or python:"
read lang
echo "got: $lang"
echo "Enter your initials: "

read initials
echo "got: $initials"
# could give them as cmd line args

# openai key export and google auth login need to be  performed before start
# modify `config.py` for lang first-before running this script

# follow terminal to interact with app
# initdb and for each lang and each user (every time we each run annotate.sh)

flask --app vec_search init-db
flask --app vec_search run --debug

flask --app vec_search export-rad-to-csv rad-$lang-lang-$initials.csv

for provider in aws openai gemini;
do
  echo "Generating relevances for: $provider";
  flask --app vec_search gen-llm-rels rad-$lang-lang-$initials.csv llm_gen_rel-$provider-$lang-$initials.csv $provider
  flask --app vec_search gen-ir-metrics llm_gen_rel-$provider-$lang-$initials.csv > metrics-$provider-$lang-$initials.txt
done;