APIs and batch processing ‐ second collaboratory session

PRE-READING

Please read Jacob Forward's wiki entry on the Anthropic workbench and his introduction to APIs

---

ACCESS OUR SESSION TWO COLAB NOTEBOOK

Please explore our session two colab notebook

[THIS LINK HAS BEEN UPDATED TO LINK TO THE ACTUAL COLAB NOTEBOOK USED IN SESSION TWO. IF YOU COME ACROSS THIS LINK AND ARE NOT A MEMBER OF TEH AI-AND-HISTORY-COLLABORATAORY BUT WOULD LIKE ACCESS TO THE NOTEBOOK, CONTACT COLIN GREENSTREET, AUTHOR OF THE NOTEBOOK]

Come to our Tuesday session with lots of questions

If you have time, try using the Anthropic API from your own Colab notebook to some some historical research on your own use cases

---

AGENDA FOR OUR SECOND SESSION

0. Feedback since first session

Gavin Beinart-Smollan: prompts to correct raw HTR Yiddish

Maurice Brenner: narrative summarization: methodological issues

1. What is an API?

2. Introducting Google's Colab Notebook

3. Example: The UK National Archives's Discovery API

Geolocating and mapping metadata downloaded from the Discovery API

4. Example: The Anthropic API

Obtaining an API and inserting it into a Google Colab notebook

Running code for a prompt in a Google Colab notebook to perform a series of historical research functions

5. API keys for other large language models

6. Use case: Using an API to process and interrogate American Presidential rhetoric

7. Use case: Using an API to process and interrogate English High Court of Admiralty depositions

8. Discussion of use cases that other collaboratory members have, which may be suited to batch processing using the Anthropic API