- daily digest of
- topic
- specific site
- questionnaire on digested content
- keeping context of last digest
- Context of n previous digests?
- ask for a topic
- limited to x count
- Inform to make it concise and clear enough for better search results
- search and keep results
- Do you want to keep an eye over x specific site or want a digest?
- yes
- Site not found in results, ask for site URL to narrow down search results
- re search and keep results
- At this point we should inform that the URL should not change, i.e. it should be a permalink (a profile link?)
- Assume it's a feed and make the source a feed source
- no
- Here we just need to search internet, take latest results and run them through content runners.
- Ask for how many pages we should look at
- Is it news? Inform it yields better results
- Ask for time of delivery
- Save this as a Task.
- Toggle active/inactive task
- Accessible via URL
- allow asking follow up questions
- Embed vectors related to question
- Show related runs <- that collected information for this digest
- show related task
- digest follow up questions history
- Task runs and their child source runs
- status of runs
- re-run a task
- Storage in SQLite
- Tasks
- topic
- related source | nil
- active
- delivery time
- delivery route
- more?
- Sources <- only persisted sources
- link
- kind
- news
- search
- feed
- meta
- Source_Runs
- <- parent task run
- <- related source
- persisted
- dynamic <- point in time storage
- status
- running
- errored
- ran
- digesting
- vectorized (think)
- done
- related digest
- Sha of this run so that we can compare if the data is still same
- think about this
- Task_Runs
- <- related task
- digest
- status
- running
- errored
- ran
- digesting
- done
- Digest
- content
- related task
- related run
- meta
- Tasks