- added first version of a user system (gf_identity)
- adding first version of p2p system based on libp2p
- initial version of gf_home, personal control panel for users
- gf_admin added, for admin login and control of a particular GF server/node
- new gf_solo service has been introduced, that compiles all other services into a single binary.
- first additions of ML functionality. Addition of Rust code for collage composition and efficient building of .tfrecords
from images. addition of basic ML models in Py.
- various improvements to gf_core, around error handling, etc. improvements to gf_rpc_lib.
- OPEN_SOURCING!! GF has been open sourced. gf_publisher, gf_landing_page, gf_tagger have also been refactored to use the new error handling structure,
thats in place for gf_images and gf_analytics (gf_crawl_lib).
- first prototype of the most basic image_editor (image filters/cropping) added. Still needs polish and integration of the UI into the
rest of the system.
- big refactor of how statistics are structured in all services, to standardize on gf_stats accross the whole system.
all stats are now accessible via gf_analytics only, not in custom dashboards of each of the services themselves
(this way expensive stats calculations are only run in the gf_analytics service machines, not on machines of other services
that are expected to efficiently handle real-time requests).
- big refactor of error handling, a system wide gf_error code is now used. not all packages migrated yet,
but core ones (gf_images,gf_crawl_lib,gf_analytics) are.
- small improvements in adding crawled_images to flows in the crawler dashboard.
- gf_crawl can move images discovered in HTML pages into image flows. This is done via UI's in the crawler
dashboard at the moment.
- first round of GIF viewer/posting features. GIF's are now clickable/playable in the flows_browser, and can
be added via the gf_chrome_ext.
- migrated system to multi-node cluster, separating key services to their own nodes. new cloud provider (GCP).
- addition of the "gf_crawl" application, which crawls the image-web for images from select sites.
these images are then accessible in their own crawled/discovered image flow.
"gf_crawl" functionality is working as a part of the gf_analytics service for now.
- addition of the "gf_domains" application. UI for it is publicly accessible, which for now gives basic stats
an all domains discovered in lings in the GF DB.
- addition of the "flows" functionality to the "gf_images" application.
- rewrite of the entire backend in Golang, and the front-end in Typescript.
major effort, and took over a year (if not longer) to complete, due to real-life coming in the way and
tech experimentation happening.
- rewrote all services and most of the front-end in Dart, very few minor improvements added.
- initial Python backend and gf_image/gf_post data-model, using mongodb, most basic css styling.
first introduction of a Chrome browser extension, for creating posts only (by adding images to them), no image flows.