MyShelf v5 Release Notes
Version 5 is packed with cool new features and improvements to make your experience even better.
What's New
Enhanced GitHub Integration
-
Direct Raw File Access
- Now, when performing GET operations like retrieving the
data.jsonfile, we use the raw files directly from GitHub. This means more stable and accurate file retrieval from your GitHub repo!
- Now, when performing GET operations like retrieving the
-
Enriched Context Sessions
- Save to Private GitHub: You can now save your enriched context session files to your private GitHub repository. All saved sessions are timestamped for easy tracking.
- Retrieve Sessions: Fetch enriched context sessions by specific dates, date ranges, or a series of dates. You can unify them into a single enriched narrative, making it a breeze to continue conversations from weeks or even a month ago.
Streamlined Workflows
-
Unified Data Workflow
- We've introduced a single workflow called 'data workflow' that manages all processing operations and cleanups. It's all about simplifying your workflow management!
-
Auto-Purging Adjustments
- Customize environment variables for auto-purging archives to keep things neat. Note: Context session files won't be auto-purged, so you can refer back to them whenever you need. Just use your best judgment on how far back you'd like to keep them.
Validation Improvements
-
YAML Validator Workflows
- Handy for those customizing YAML workflows. Ensures your configurations are spot-on and error-free.
-
JSON Schema & Template Validations
- Prevents incorrect changes from messing with your root
data.jsonfile. Plus, all new inbounddata.jsonfiles are archived to keep your data safe.
- Prevents incorrect changes from messing with your root
-
Supportive Python Scripts
- Added several Python scripts in the root path to help maintain schema and template adherence.
Better Management Features
-
Personas Management
- Centralized Storage: No more attaching personas to custom GPTs, except for the default mode and persona mode template. All personas are now stored in a dedicated
personasfolder in your private GitHub repo. - New Commands: Use the new
switchmode {mode}command to easily switch between different modes.
- Centralized Storage: No more attaching personas to custom GPTs, except for the default mode and persona mode template. All personas are now stored in a dedicated
-
Recipes Folder
- A new
recipesfolder in the GitHub repo organizes indexes for breakfast, lunch, dinner, and even cocktails. This works seamlessly with the "MyBar" feature indata.json.
- A new
-
Proto-Cognitive Graph
- Introducing a proto-cognitive graph of your filesystem in your private GitHub repo.
Important Notes
- Personas & Recipes Folders
- These folders are significant upgrades from storing everything in
data.json, which previously caused performance issues as the file grew too large. Now, with dedicated folders containing aindex.jsonfiles further complimented by graphs, you can manage your data needs more efficiently. This setup opens up a wealth of opportunities for better data management and scalability.
- These folders are significant upgrades from storing everything in
Thanks for being a part of the MyShelf community! We hope you enjoy these new features and improvements. As always, feel free to reach out with any feedback or questions.
Happy Shelving! 📚✨