💾 Archived View for thebird.nl › gn-gemtext-threads › topics › biohackathon › GNGSoC2023.gmi captured on 2023-09-08 at 16:38:04. Gemini links have been rewritten to link to archived content
View Raw
More Information
⬅️ Previous capture (2023-07-22)
🚧 View Differences
-=-=-=-=-=-=-
GNSoC 2023
GN Summer of Code
Introduction
We are running a GN Summer of Code in small teams.
- Runs July + August 2023
- Weekly plenary where projects present progress - Thursday 9am EU, 10 am EAT.
- Projects should be (slightly) out of comfort zone
- Use gemtext documentation
- Option for publishing on progress by then end as a BLOG or BioHackrXiv
CI for guix-bioinformatics (guix pull)
- lead: Arun
- team: Efraim, Pjotr, Sarthak
Making GN deployment rock solid
git repo genenetwork-machines, guix-bioinformatics
Tracking progress
Week 1
- guix pull on guix-bioinformatics
- updated guix is broken gemma (Pjotr)
- Efraim guix GN2 - so we can have a channel
- Next step build substitutes for guix-bioinformatics
- once built they are shared
- And create a GN3 channel (Efraim?)
Week 2
- RISC-V port progressing with node and zig 0.10
- guix-bioinformatics now has CI!
- ~700 packages, 240 are broken ;)
Week 3
- Arun gives a presentation on laminar using guix-forge: slides:
https://issues.genenetwork.org/topics/ci-rethink-slides
- New system is simpler and has reproducibility issues
- Efraim is doing channels for GN2 and GN3
- Sarthak will try to run guix-forge
Week 4
- Arun added channels to CI
- localhost
- cgit
- Efraim: gn2 -> channel; Arun tested
Week 5
- Added Klaus server for git
- Fixed channels to work wit Python 3.9 instead of 3.10
Week 6
- Towards a new workbench with cgit support
Week 7
- Arun has cgit running on his own server - soon bringing up tux02 after resolving https
Week 8
- cgit deployment at https://git.genenetwork.org/
- guix forge
Week 9
- Discussion on importers for guix forge
- Talked about propagator networks - part of Seattle presentation
More
https://ci.genenetwork.org/jobs/guix-bioinformatics
- CI/CD is up and running again (and broken)
- Rethink: channels and pull channels are used for CI/CD
- Move unused packages elsewhere
Nextgen databases
lmdb+RDF
- lead: Bonface
- team: Fred, Alex
- contact: Pjotr
git repo genenetwork3
Design doc
Week 1
- RDF dumps
- Parsing S-exp -> markdown
- Hashing tables (Fred)
- automated updates
- Some progress on sample data from SQL -> lmdb (Alex)
- Next week: guile bindings for lmdb
- improving RDF
Week 2
- RDF structure to markdown dump
- Fred is running SPARQL queries
- Alex is adding lmdb phenotype API endpoints
Week 3
- Bonface demoes new documentation & code
Week 4
- Settled on prefixes terms and id
- Updated man pages
https://github.com/genenetwork/gn-docs/tree/master/rdf-documentation
- Started work on lmdb+guile:
https://github.com/BonfaceKilz/gn-data-vault
Week 5
- Metadata - renamed prefixes
- Short names gn: gnt:
- Updated virtuoso
- parsing geno files - lmdb support
Week 6
Week 7
- RDF improvements with ontology
- Inconsistencies and privacy discussion today
Week 8
- Transformed most tables now
- guile-lmdb started by Alex
Week 9
- New renaming and modelling
- GeneRif
- Working on unique IDs
- SKOS
LLMs & metadata (RDF)
- lead: Shelby
- team: Priscilla
- contact: Bonface, Pjotr, Rupert
Tracking progress
Week 1
- Created issue page
- Downloading publications (Priscilla)
- Flask server
- Next: Connecting OpenAI
- Create matrix room
Week 2
- Open AI API is working
- Shelby is integrating into a Flask interface for GeneNetwork
- Using a pubmed UI style
Week 3
- Shelby shows code
- Plan to host
- Priscilla is working on SLA and document acquisition
- Hosting GN Q&A
Week 4
- working on container
- fetched 1000 publication
- JSON documentation on references
Week 5
- Guix container for LLM
- Expose container to Rupert
- Add a password
Week 6
- Very close to a working flask app Rupert can try next week
Week 7
- Working FLASK app ready for testing - Rupert will have a go
Week 8
- Working prototype!
- references as JSON
Week 9
- Shelby showed working references and challenges
- System is ready for viewing by Rob
API to access data from GN
- lead: Rupert
- team: Flavia
- contacts: Bonface, Zach, Fred
Documentation and adding endpoints
git repo gn-docs & genenetwork3 & SPARQL
https://github.com/genenetwork/gn-docs
Week 1
https://github.com/genenetwork/gn-docs/blob/master/api/questions-to-ask-GN.md
- Ideas on structuring
- Questions on GN
- Next: unify access to information
- collecting questions from users
- settle on form of API
- create example URLs for mouse
Week 2
- GraphQL Arun gives a mumi demo - schema allows for (partial) queries and querying the schema itself
- Pjotr convenience API demo - add endpoints in results
- Flavia added questions in gn-doc - e.g. for synteny search
Examples for synteny
Week 3
- Rupert proposes endpoints and metadata traversing
https://github.com/genenetwork/gn-docs/blob/master/api/alternative-API-structure.md
Week 4
- Start working on endpoints
- R api - reference GN/API
- synteny
Week 5
- Added back-end support for wikidata - finding inconsistencies
Week 6
- API ready for running in a production environment
- Using latest RDF
Week 7
- Test version of API is running at https://luna.genenetwork.org/api/v2.0/
- We'll continue building up facilities
Week 8
- R script to parse API by Rupert
- Move forward with populations/strains and datasets
Week 9
- Progressed API software to include groups
Editing data
- lead: Fred
- team: Arthur, Rupert, Zach
- contacts: Rob
Week 1
- Edit phenotype metadata works
- Next: phenotype values and testing on live
Week 2
- Fixing issues
- Meeting on requirements from Arthur and Zach
Week 4
- Editing works!
- Discussed the approval procedure and edit button for everyone
Week 5
- Editing has gone on production - fixing issues
- Discussion on REST API
Week 6
- Progress on authorization and editing
Week 7
- Fred has moved code into a new repo for https://github.com/genenetwork/gn-auth
Week 8
- gn-auth is building locally and needs to go on the forge
- case attributes
Week 9
- gn-auth continuation
- case-attribute editing progress and challenges
Guix parametrization
- lead: Sarthak
- team: Pjotr, Gabor
- contacts: Ludo
Week 1
https://blog.lispy.tech/parameterized-packages-an-update.html
- Next: focus on statically built packages optimized for arch.
Week 2
- Looking into GeneNetwork3 service
- Enumerated types
Week 3
- Preparing for BLOG on S-exp
Week 4
- BLOG in a stage that we discuss naming conventions
Week 5
- Proposed DSL for parameters
Week 6
- Posted BLOG and started implementation
Week 7
- Agreed on final deliverables for GSoC
Week 8
- Milestone on dependencies!
Week 9
- Sarthak showed us the working prototype
Links
Arun's talk on our issue tracker
Git repo on issues/tasks/topics
This page
For more info contact pjotr.public912 at thebird.nl