v0.5.12 Release Notes

We are excited to announce the stable release of v0.5.12 for the Obsidian TextGenerator Plugin! This version brings several key enhancements, bug fixes, and new features designed to improve your experience and increase flexibility.

Highlights of v0.5.12:

• Support for gpt-4-1106-preview (gpt4 turbo): The plugin now supports the latest GPT-4 model with turbo capabilities, offering more powerful and efficient text generation within Obsidian.

• Extensive Language Model Compatibility: Enjoy flexibility with support for a diverse range of language models, including gpt3.5, gpt4, gpt-3.5-instruct, claude, bard, and llama using Langchain.

• Add Playground: We've refined the playground UI to create a more visually appealing and user-friendly experience when crafting prompts and viewing responses.

• Better Settings: Visually appealing and user-friendly experience for the settings with search.

• Advanced Variable Access: All variables are now accessible within custom Instruct templates, providing you with greater control over dynamic content generation.

• Advanced Template Engine: Elevate your productivity with dynamic Templates (Handlebars & Javascript) with chaining that simplify and enhance your note-taking routine.

• Open Template as Tool : Open Template as Tool is an innovative solution designed to streamline your workflow by integrating templates directly into your desktop environment. This tool ensures your templates are accessible and editable in real time, atop any application you might be using.

• New Helpers for Enhanced Processing:

  • encodeURI helper for better URI encoding.
  • regex helper introduced for sophisticated text manipulation.
  • Each helper updated to handle objects as well as arrays.

• Efficient Data Handling: Enhanced variable extractions using Handlebars templates make data management more streamlined and powerful.

• Optimizations and Bug Fixes: A range of optimizations have been implemented to ensure smoother performance, alongside numerous bug fixes that address issues from previous versions.

Other Notable Changes:

• Default setting for "free cursor" is now false, enhancing user experience by avoiding unexpected cursor movements.

• Improved slow-loading page handling, with specific attention to platforms like Twitter and Reddit.

• Deprecated the term "engine" in favor of "model" across documentation and settings for better clarity.

• New read, write, and append helpers have been added to extend functionality.

• Introduced run and extract functions in the script helper for a more streamlined approach.

• The starting template example has been enhanced to provide clearer guidance for new users.

• Fixed issues with relative links during web extractions ensuring accurate data retrieval.

New Documentation Available

Alongside this release, we're also thrilled to introduce our newly revamped documentation available at https://docs.text-gen.com/. Here you can find comprehensive guides, usage tips, and best practices that will help you leverage the full potential of the Obsidian TextGenerator Plugin.

We hope you enjoy these updates as much as we enjoyed developing them! Your feedback is always welcome as it helps us continue improving your writing workflow within Obsidian.

Happy text generating!

Release v0.5.12-beta 🚀

Changes:

• Fixed documentation link 📚

Release v0.5.11-beta 🚀

Changes:

• Fixed mobile bugs 📱
• Fixed minor bugs 🐞

Release v0.5.8-beta 🚀

Changes

• Fixed the extractor tool for RSS feeds for improved functionality 🌐
• Resolved the bug in template tool variables for smoother operation 🛠️
• Rectified a strange bug with the package manager for better performance 🐞
• Updated the documentation link for the latest resources 📚
• Modified the getRandomFile helper to utilize regex for more accurate results 🔀
• Fixed processing with sub-templates for streamlined template handling 🔄
• Corrected issues related to variables for consistent and predictable behavior 🐛
• Addressed the set helper bug for more reliable variable setting 🛠️
• Handled the bug occurring when there are no variables present for smoother execution 🐞
• Fixed the "read" helper for better data retrieval and handling 📚
• Integrated an RSS parser for improved RSS feed management 📰
• Addressed a bug with the gen function in the script for better script execution 🛠️
• Added "yt" as a slug for YouTube for enhanced linking 🌐
• Implemented "gen" in script helper function for streamlined data generation 🔄
• Set to ignore anything inside the script block for improved script management 📜
• Rectified the bug where adding space in string results caused issues in added variable barhandles 🐛
• Fixed a startup issue with the llm Provider for smoother initiation 🚀
• Addressed an inconsistency bug with the run helper for more predictable outcomes 🐞
• Resolved various internal bugs for enhanced stability and performance 🐞

Release v0.5.8-beta 🚀

Changes:

• Preparing for the stable release 🏗️
• Set "free cursor" to default to false for improved user experience 🔒
• Fixed extraction issues that occurred with slow-loading pages like Twitter and Reddit 🐢
• Resolved the issue with "make template" from the playground 🛠️
• Removed the "eachProperty" and updated "each" to also work with objects 🔄
• Improved version upgrading and ensured backward compatibility 🔄
• Deprecated the word "engine" and replaced it with "model" for clarity ⚙️
• Added read, write, and append helpers for enhanced functionality ✏️
• Introduced run and extract functions to the script helper for streamlined usage 📜
• Added a link to Discord for more templates/packages 🌐
• Changed "VAR_" to "vars" for improved naming consistency 🔄
• Enhanced the starting template example for better understanding 📝
• Fixed relative links when extracting from the web 🌐
• Numerous internal bug fixes for smoother performance 🐞

Full Changelog: https://github.com/nhaouari/obsidian-textgenerator-plugin/compare/0.5.7-beta...0.5.8-beta

Release v0.5.7-beta 🚀

Changes:

• Fixed the playground UI for improved user experience 🎨

• Added yaml and metadata variables for more comprehensive data handling 📄
  

• Made all variables accessible within the custom Instruct template for increased flexibility ✨

• Added encodeURI and eachProperty helpers for better URI handling 🌐

• Introduced a regex helper to handlebars for improved text processing 🧹

• Enhanced variable extractions with handlebars for more efficient data handling 💡

• Implemented several optimizations and bug fixes for smoother performance 🐞

Full Changelog: https://github.com/nhaouari/obsidian-textgenerator-plugin/compare/0.5.6-beta...0.5.7-beta

Release v0.5.6-beta 🚀

Changes:

• Added the window playground handlebars to preview how the prompt will look like 🎨
  you can access it by enabling open-playground command.
  

• Changed selection to tg_selection, and reserved the variable selection solely to get the highlighted/selected text 🔄

• Added custom instruct + {{context}} template in the settings page, allowing for more customizations ✏️
  
  

• Included the http-Referer header as suggested by pipboyguy#0000 to work with openrouter's endpoint

• Attempted fix for "Disable auto-suggestions at the beginning of the line #200" 🐞

Full Changelog: https://github.com/nhaouari/obsidian-textgenerator-plugin/compare/0.5.5-beta...0.5.6-beta

Release v0.5.5-beta 🚀

Changes:

• Corrected the issue where it took context instead of selection in normal generation 🐞
• Made minor fixes in the package manager for improved performance 🛠️
• Added slugs to providers for better organization and access 🌐
• Resolved the problem with undefined URL in the audio extractor 🎧

Full Changelog: https://github.com/nhaouari/obsidian-textgenerator-plugin/compare/0.5.4-beta...0.5.5-beta

Release v0.5.4-beta 🚀

Changes:

• Added support for the plugin API 🧩

Plugin.tg.gen("hello").then(console.log)

• Eliminated unnecessary console logs for smoother operation 🔄

• Fixed a bug that was preventing generation 🐛

• Rectified the scroll issue in the output of the tool, especially when using a theme 🛠️

• Enhanced the tool's user interface for improved usability and aesthetics 🎨

• Fixed the source template for better functionality ✅

Full Changelog: https://github.com/nhaouari/obsidian-textgenerator-plugin/compare/0.5.3-beta...0.5.4-beta

Release v0.5.3-beta 🚀

Changes:

• Updated the default behavior of the tool to create a new popup instead of a side view 🔄
• Implemented a version manager to handle breaking changes between versions and ensure backward compatibility 🛠️
• Made minor adjustments and improvements for enhanced performance and user experience ✨

Full Changelog: https://github.com/nhaouari/obsidian-textgenerator-plugin/compare/0.5.2-beta...0.5.3-beta

Release v0.5.2-beta 🚀

Changes:

• Added the ability to set a custom variable name using the format {#run "templateId" "customVarName"} (optional but helpful in some cases) 🛠️
• Added "open as tool" functionality to templates, along with new commands "open template as tool" and "tool" which can be set in any template 🧰
• Added a popout and "always on top" feature to the tool 🪄
• Numerous bug fixes 🐞

changes:

• Fixed mobile support, by forking library handlebars-async-helpers and removing readable call

Version 0.5.0-beta

Changes:

• Template calling other templates
• Template inheritance
• Batch generate using template on multiple files (was highly requested)
• Template calling extractors ex:{{extract "web" url}}
• Langchain chains support (this will allow a huge amount of text to be summarized) setting chain.type: "map-reduce" will enable it
• Bug fixes

Additional Information:

• Examples of templates are attached as examples.zip. These examples showcase the power of calling templates from other templates.

Release Notes - Version 0.4.9-beta

Bug Fix

• Fixed Models List Not Showing Up: This release addresses a critical issue where the models list was not displaying as expected.

We appreciate your feedback and contributions to make this project even better. If you encounter any issues or have suggestions for improvement, please open an issue or submit a pull request.

Thank you for using our plugin!

Release Notes - Version 0.4.8-beta

New Features

• Added Summary Funtionality with langchain:

template:

---
promptId: Summarize
name: Summarize
title: Summarize long text
description: Summarize long text
commands:
  - generate
mode: insert
chain.type: map_reduce
splitter.ChunkSize: 2000
splitter.ChunkOverlap: 200
---

{{selection}}


***
Summary: {{output}}

Bug Fixes

• Fixed Switching Between Encryption Enabled/Disabled: A bug that prevented smooth switching between encryption enabled and disabled modes has been resolved.

• Fixed Free Cursor Not Working: A recent bug affecting the free cursor functionality has been addressed and is now functioning as expected.

Enhancements

• Changed Naming of LLM Providers: We have updated the naming conventions for LLM (Language Model) providers for clarity and consistency.

Thank you for using our software!

Release Notes

Changes

• Fixed mobile support.

• Improved error handling with better error messages if text generation fails.

Release Notes

Version 0.4.6-beta

Changes

• Improved Text Selection/Insertion Logic:

  • Previously, the tool selected the entire block (line or document) before making changes. Now, it will only select the text before the cursor position, making editing more precise.
  • Text will now be inserted in the same position as the cursor, rather than at the end of the block.

• Enhanced Context Display:

  • The context (selection) will now be shown before sending, providing better context-awareness for users.

• Provider Selection in Frontmatter:

  • Added a new feature that allows users to select a Provider within the frontmatter, enhancing customization options.

• Resolved Section Height Issue:

  • Fixed a bug that caused abnormal height behavior in the settings section, improving the overall user experience.

• Various Bug Fixes:

  • Addressed several bugs and issues to ensure smoother functionality and reliability.

We appreciate your feedback and support <3

Release Notes - Version 0.4.5-beta

Changes:

• Auto Suggest Fix: Fixed issues, now working as intended.
• Model Selection: Updates selected model accurately.
• Status Bar: Debounced for maxTokens changes.

Enjoy!

Release Notes - Version 0.4.4-beta

Changes:

• OpenAI LLMs Support: Added support for OpenAI's latest language models, including the new gpt-3.5-turbo-instruct, enhancing the capabilities of our platform and providing users with even more powerful AI assistance.

• Updated Models List: We've updated the models list to ensure that you have access to the latest and most relevant AI models, keeping our platform up-to-date with the latest advancements in AI technology.

• Automatic Deployment: Implemented automated deployment for both stable and beta versions of the plugin using GitHub Actions. This streamlines the release process, ensuring that users can quickly access the latest features and improvements.

• Command Loading Bug Fixes: Resolved several bugs related to loading commands when the plugin reloads. Users will now experience a smoother and more reliable operation when interacting with the plugin.

• Settings Page Resizing: Fixed bugs related to resizing the settings page, preventing content from being hidden or obscured. This improvement enhances the user experience, making it easier to configure and customize settings.

Thank you for using our plugin, and we hope you enjoy the enhanced features and improvements in this release!