gpt-3.5-turbo, priced at $0.002 per 1k tokens (#21)

The current formula for cost calculation is not correct for the current pricing for gpt-3.5-turbo so we need to add a zero here to correct the calculation from .02 to .002 per 1000 tokens.    https://openai.com/pricing

Author: scott------

paperqa/docs.py

@@ -361,7 +361,7 @@ def _query(
         formatted_answer = f"Question: {query}\n\n{answer_text}\n"
         if len(bib) > 0:
             formatted_answer += f"\nReferences\n\n{bib_str}\n"
-        formatted_answer += f"\nTokens Used: {tokens} Cost: ${tokens/1000 * 0.02:.2f}"
+        formatted_answer += f"\nTokens Used: {tokens} Cost: ${tokens/1000 * 0.002:.2f}"
         answer.answer = answer_text
         answer.formatted_answer = formatted_answer
         answer.references = bib_str