Create a browser bookmark with the following code as the URL, name it Slack Timesheet
or something
javascript:(async function() {
allMessages = [];
while (true) {
buttons = document.getElementsByClassName('c-search__expand');
buttons = Array.from(buttons);
for (b of buttons) { b.click(); }
pageMessages = document.getElementsByClassName('p-rich_text_section');
pageMessages = Array.from(pageMessages).map(m => m.innerText.replace('\n', ' ... '));
allMessages = allMessages.concat(pageMessages);
button = document.querySelector('[data-qa=c-pagination_forward_btn]');
if (!button) { break }
button.click();
await new Promise(r => setTimeout(r, 500))
}
text = allMessages.join(' ... ');
instruction = [
'Act like a personal assistant',
'I need a timesheet',
'You will determine the tasks I did using the chat messages that I sent at work',
'In the process, you should write 1 line of description for each task, but only consider the 5 most important tasks if there are more than 5',
'Please do not exceed 14 words for each task, and avoid any output other than the 1 line description for each task',
'The chat messages that I sent will follow',
].join('. ');
prompt('Copy this', `${instruction} >>> ${text}`)
})()
Go to Slack, search for all messages that you sent on a certain date, for example:
from:@Phu before:2024-02-24 after:2024-02-22
Stay on the first result page, click the bookmarklet to run it. It will:
- Expand long messages
- Go to all result pages to get all messages
- Build a prompt suitable for LLMs like ChatGPT and present it so you can copy
Copy the prompt and paste it at e.g. https://chat.openai.com/
The result looks something like this
1. Analyzed autovacuum performance, proposing optimization strategies.
2. Investigated session limit and CPU usage for database operations.
3. Prioritized DB migration ticket DO-3512 for execution.
4. Identified missing database migration in pay_staging-policies-137.
5. Discussed data export/import plan with concerns about permissions.
Touch up as needed