enhance: add document summarizer example

[g-linville]

No description provided.

[cjellick]

Testing this out now

[cjellick]

hit this:

14:16:00 started  [main]
14:16:00 sent     [main]
         content  [1] content | Waiting for model response...
         content  [1] content | tool call hamlet-summarizer -> {"index":"0"}
14:16:02 started  [hamlet-summarizer(2)] [input={"index":"0"}]
14:16:02 sent     [hamlet-summarizer(2)]
         content  [2] content | Waiting for model response...
         content  [2] content | tool call hamlet-retriever -> {"index": "0"}
         content  [2] content | tool call sys-read-4381effb -> {"filename": "summary.txt"}
14:16:05 started  [hamlet-summarizer(3)->sys.read(3)] [input={"filename": "summary.txt"}]
14:16:05 started  [hamlet-summarizer(4)->hamlet-retriever(4)] [input={"index": "0"}]
14:16:05 sent     [hamlet-summarizer(3)->sys.read(3)]
14:16:05 failed to run tool [hamlet-retriever] cmd [python3 main.py 0]: context canceled
14:16:05 sent     [hamlet-summarizer(4)->hamlet-retriever(4)]
2024/03/01 14:16:05 open summary.txt: no such file or directory

[cjellick]

• Make it so that we dont have to touch the summary file to start
• I think tool.gpt should be reserved for ...reusable tools maybe? Your script is so specific that i think it should just be called hamlet-summarizer.gpt
  is around to agree or disagree on this)
• Id like the readme to give a little more detail on setup. like, you need to install requirements.txt. ...i also think we should start including the the virtualenv step in these instructions

[g-linville]

@cjellick I just pushed a change for all three of those. Creating the summary.txt file in the Python script didn't work, since the summarizer tool will still attempt to read it before it even gets its first piece of the text. So I just have the top-level tool create it before it starts doing anything else.