71d4581110
refactor(ai-client): moving tool handling and client into seperate folders
2025-04-04 22:03:46 +01:00
8a165c2042
wip(orchestrator): basic scaffolding for the agent
2025-04-04 20:40:31 +01:00
fe7c92b622
feat: sample data
2025-04-02 17:32:52 +00:00
745265773b
feat(notes): allowing frontend to save
2025-04-01 20:54:15 +00:00
f72ee73020
feat(notes): saving the notes for any images for easy text searching
2025-04-01 20:45:43 +00:00
1cb6510465
feat(events): search through organizer
2025-04-01 19:59:17 +00:00
9652549b01
fix: adding location to general query
2025-04-01 19:32:55 +00:00
901f214f9d
feat(contacts): events can now have organizers
2025-03-31 18:40:36 +00:00
0c78d741f0
feat(schema): basic contact tables
2025-03-31 18:10:04 +00:00
13e82334ca
feat(events): adding start and end times
2025-03-31 17:32:55 +00:00
59737ca9ac
feat(schema): removing coordinates and adding start times to events
...
.
2025-03-31 16:44:42 +00:00
b6969127eb
fix(backend): SQL statements without returns
2025-03-26 16:51:46 +00:00
d716e66463
feat: attaching both to image
2025-03-26 16:22:23 +00:00
a576355e7c
feat: creating events and attaching locations
2025-03-26 16:16:48 +00:00
6f938a34e3
feat(tool-calling) Big refactor on how tool calling is handled
...
these commits are too big
2025-03-22 20:46:26 +00:00
7debe6bab2
feat(locations): allowing AI to attach it to the image
2025-03-22 17:47:02 +00:00
4c4bf7a9e4
feat(location): working e2e with tool calling
2025-03-22 12:22:31 +00:00
aad45fcf52
feat(tool-calls): listLocation tool call handling
2025-03-22 11:14:00 +00:00
7c473e054a
feat: using tools for event loocation agent
2025-03-22 10:12:51 +00:00
7f96d2fc45
refactor(naming): using Agent instead of openai
2025-03-21 17:07:00 +00:00
df90eaa6ad
fix: docker image
2025-03-21 14:49:51 +00:00
44d506bc69
wip(frontend): adding more information
2025-03-21 14:36:03 +00:00
58f7afb521
fix: re-adding location to event
...
Figured it out
2025-03-21 14:08:06 +00:00
7b64563647
fix: simplifying query
...
I do have some learning to do about go-jet specifically.
I can't include the Location field on events, because the QRM will add
any location to the events
2025-03-21 14:01:21 +00:00
72a3e58ef9
feat: returning events and locations from end point
2025-03-20 18:34:11 +00:00
f042c9dfcc
feat: working e2e solution
2025-03-20 17:59:00 +00:00
d2dd43c6b2
merging
2025-03-19 09:46:52 +00:00
56d423d261
feat: adding events and locations to json schema
2025-03-19 09:46:42 +00:00
072eebc0bf
feat(events): also working
2025-03-18 18:37:12 +00:00
28dd02a47d
feat(locations): now working
2025-03-18 18:18:01 +00:00
2e1809aa27
refactor(text,tags,links): to foreign key to image instead of user_image
2025-03-18 17:48:38 +00:00
9b5113f428
Merge branch 'main' of https://git.johncosta.tech/JohnCosta27/Haystack
2025-03-17 21:01:15 +01:00
b29a013cde
wip
2025-03-17 21:00:52 +01:00
439f729150
fix: returning whole tag object
2025-03-16 18:29:15 +00:00
bfd6d136dc
fix: returns correct image ID
2025-03-16 18:20:29 +00:00
1d5d90c3b5
refactor(models): using more organised structure
2025-03-16 18:13:30 +00:00
d4c6aa0310
feat(tags): correctly inserting new tags and adding them to images
2025-03-11 22:47:28 +00:00
6fcb1e9f26
feat(tags): creating and getting user tags
2025-03-11 21:23:41 +00:00
c215bf6909
feat: new schema to support user tags better
2025-03-11 20:29:56 +00:00
9e3896a30f
feat: super basic image search
2025-03-08 15:37:10 +00:00
e5ac7061f4
feat: sending images and receiving them is now working
2025-03-08 13:13:05 +00:00
03a4d49ee6
feat: sending base64 image to backend
...
This is silly, but binary is apparently hard to do????
2025-03-08 12:30:16 +00:00
53c4ec1869
fix: using json response header
2025-03-07 14:14:40 +00:00
5192aeb70f
wip: Using mistral instead of OpenAi
2025-03-07 13:42:50 +00:00
a3dee09011
feat: making prompt be more generic with tags
2025-02-26 20:53:12 +00:00
a4136fe3f8
feat: using different prompt
2025-02-26 20:49:55 +00:00
0ec2ae65d2
fix: some spam from get images request
2025-02-26 20:48:52 +00:00
79ee5e23ca
fix: getting user images
2025-02-26 20:09:19 +00:00
1f16cfb30b
refactor: tables for image and processing_image
...
This allows a single table to be used to process images, meaning if
anything happens to the system we can always return to polling the
database and process these images individually.
Because of this we also want an `image` table to contain the actual
binary data for the image, so we aren't selecting and writing it each
time, as it is potentially a bottleneck.
2025-02-26 20:01:56 +00:00
2a37e37c4b
refactor: using chi router + bug fixes
2025-02-26 18:04:30 +00:00
