Improved directory initialization, template management, and temporary file handling.

[feat] Added a feature to copy templates into $GIV_HOME/templates directory
[fix] Resolved an issue where templates were not properly copied to user's project directory
[chore] Improved script organization by moving common tasks to new system.sh helper functions

Author: itlackey

.gitignore

@@ -13,4 +13,6 @@ dist
 .giv/tmp
 .giv/cache
 .giv/logs
-tests/.giv
+tests/.giv
+.notes
+.giv/config.bak

.giv/config

@@ -1,10 +1,39 @@
-GIV_TMPDIR_SAVE=1
+# GIV Configuration Example
+## Check the docs/configuration.md for more details.
 
-### Basic setup for Ollama CLI
-GIV_MODEL=devstral
-# Note: When using remote generation, the GIV_API_MODEL variable or the CLI --api-model arg will determine which model to use
+## Debugging
+# ### Set to "true" to enable debug mode, which provides detailed logging
+# GIV_DEBUG=""
+# ### If set to "true", enables dry run mode (no changes will be made)
+# GIV_DRY_RUN=""
+# ### Set to "true" to preserve temporary directories after execution for debugging purposes
+# GIV_TMPDIR_SAVE=""
+
+
+## Model and API configuration
+GIV_MODEL_MODE="auto"
+GIV_MODEL="devstral"
 
 ## Remote: groq API Configuration
 GIV_API_URL="https://api.groq.com/openai/v1/chat/completions"
 GIV_API_MODEL="compound-beta"
-GIV_API_KEY="${GROQ_API_KEY:-}"
+GIV_API_KEY="${GROQ_API_KEY:-}"
+
+
+## Remote: OpenAI Compatible API Configuration
+GIV_API_MODEL=gpt-4o-mini
+GIV_API_URL=https://api.openai.com/v1/chat/completions
+GIV_API_KEY="${OPENAI_API_KEY:-}"
+
+## Project details
+### Path to the file containing the project version (e.g., "pyproject.toml" or "package.json")
+GIV_VERSION_FILE="src/config.sh"
+
+### Regex pattern to extract the version string from the version file (e.g., 'version\s*=\s*"([0-9\.]+)"')
+GIV_VERSION_PATTERN=""
+
+### Regex pattern to identify TODO comments in code (e.g., 'TODO:(.*)')
+GIV_TODO_PATTERN=""
+
+### Comma-separated list of files or glob patterns to search for TODOs (e.g., "*.py,src/**/*.js")
+GIV_TODO_FILES="docs/todos.md"

docs/using_changeish_with_npm.md docs/articles/using_giv_with_npm.mddocs/using_changeish_with_npm.md renamed to docs/articles/using_giv_with_npm.md

@@ -0,0 +1,36 @@
+# GIV Configuration Example
+## Check the docs/configuration.md for more details.
+
+## Debugging
+### Set to "true" to enable debug mode, which provides detailed logging
+GIV_DEBUG=""
+### If set to "true", enables dry run mode (no changes will be made)
+GIV_DRY_RUN=""
+### Set to "true" to preserve temporary directories after execution for debugging purposes
+GIV_TMPDIR_SAVE=""
+
+## Default git revision and pathspec
+### GIV_REVISION="--current" uses the currently checked-out git branch or commit.
+GIV_REVISION="--current"
+GIV_PATHSPEC=""
+
+## Model and API configuration
+GIV_MODEL_MODE="auto"
+GIV_MODEL="devstral"
+GIV_API_MODEL=""
+GIV_API_URL=""
+### Uses the OPENAI_API_KEY environment variable if set, otherwise empty
+GIV_API_KEY="${OPENAI_API_KEY:-}"
+
+## Project details
+### Path to the file containing the project version (e.g., "pyproject.toml" or "package.json")
+GIV_VERSION_FILE=""
+
+### Regex pattern to extract the version string from the version file (e.g., 'version\s*=\s*"([0-9\.]+)"')
+GIV_VERSION_PATTERN=""
+
+### Regex pattern to identify TODO comments in code (e.g., 'TODO:(.*)')
+GIV_TODO_PATTERN=""
+
+### Comma-separated list of files or glob patterns to search for TODOs (e.g., "*.py,src/**/*.js")
+GIV_TODO_FILES=""

docs/code-review-01.md

@@ -1,43 +1,46 @@
-
 ### **Outstanding Tasks for Version 0.4.0**
 
+**Refactoring:**
+- Modularize the `summarize_target()` and `summarize_commit()` functions in `history.sh`.
+  - Ensure edge cases like invalid commits or empty diffs are handled.
 
- **Refactoring:**
-   - Simplify flow control for model invocation.
-   - Modularize the `summarize_target()` and `summarize_commit()` functions.
-
- **Configuration Enhancements:**
-   - Implement the polished `parse_args()` function from giv_folder.md.
-   - Add .giv initialization logic.
+**Testing:**
+- Expand tests for `summarize_target()` and `summarize_commit()` in `summarize_target.bats`.
+- Increase automated testing coverage for text insertion and token replacement functions.
 
- **Testing:**
-   - Increase automated testing coverage for:
-     - Text insertion and token replacement functions.
-     - The new `parse_args()` and .giv initialization logic.
+**Documentation:**
+- Update the roadmap and README to reflect completed and pending tasks.
 
 ---
 
-### **Implementation Plan**
+### **Completed Tasks**
 
-#### **Simplify Flow Control for Model Invocation**
-- **Steps:**
-  1. Refactor the logic in giv.sh to streamline model invocation.
-  2. Ensure `summarize_target()` respects `--dry-run` and skips calling the model when `--model-mode none` is set.
+**Configuration Enhancements:**
+- Implemented the polished `parse_args()` function in `args.sh`.
+- Added `.giv` initialization logic via `ensure_giv_dir_init` in `system.sh`.
+
+**Bug Fixes:**
+- Resolved the `--api-key` parsing issue in `args.sh`.
+
+**Testing:**
+- Added tests for `.giv` initialization in `test_giv_init.bats`.
+- Added tests for `replace_tokens` in `replace_tokens.bats`.
+
+---
+
+### **Implementation Plan for Remaining Tasks**
 
 #### **Modularize `summarize_target()` and `summarize_commit()`**
 - **Steps:**
-  1. Break down `summarize_target()` in history.sh into smaller, reusable functions.
-  2. Ensure `summarize_commit()` in history.sh is modular and handles edge cases like invalid commits or empty diffs.
-
+  1. Break down `summarize_target()` into smaller, reusable functions.
+  2. Refactor `summarize_commit()` to handle edge cases and improve readability.
 
-#### **Add .giv Initialization Logic**
+#### **Expand Testing Coverage**
 - **Steps:**
-  1. Add .giv initialization logic to system.sh using the `ensure_giv_dir()` function.
-  2. Implement the `ensure_giv_dir()` function in giv.sh.
-  3. Ensure it creates the .giv directory and default configuration files on the first run.
+  1. Add unit tests for modularized `summarize_target()` and `summarize_commit()`.
+  2. Expand tests for text insertion and token replacement functions.
 
-#### **6. Increase Automated Testing Coverage**
+#### **Update Documentation**
 - **Steps:**
-  1. Verify unit tests for `parse_args()` in test_parse_args.bats.
-  2. Add tests for .giv initialization logic in a new test file (e.g., `tests/test_giv_init.bats`).
-  3. Expand tests for token replacement in replace_tokens.bats.
+  1. Update the roadmap to reflect completed tasks.
+  2. Ensure README includes details of new features and fixes.

docs/config.example

@@ -10,9 +10,8 @@
   - Simplify flow control for model invocation to improve maintainability.
   - Modularize the `summarize_target()` and `summarize_commit()` functions for better readability and testing.
 - **Configuration Enhancements:**
-  - Implement the polished `parse_args()` function from the `giv_folder.md` file.
   - Add `.giv` initialization on first run with default templates and configuration.
-  - Ensure proper precedence for configuration sources: CLI args → config file → `.giv/.givrc` → environment variables.
+  - Ensure proper precedence for configuration sources: CLI args → config file → `.giv/config` → environment variables.
 - **Testing:**
   - Increase automated testing coverage for text insertion and token replacement functions.
   - Add unit tests for the new `parse_args()` and `.giv` initialization logic.

docs/giv_folder.md

@@ -1,32 +1,5 @@
 # TODOs
 
-## Git Command Robustness:
-The use of `git diff` and `git rev-list` assumes the repository is in a clean state. If the user runs giv inside a shallow clone or with unusual ref names, some commands could fail.
-
-The script does check `git rev-parse --is-inside-work-tree` early to ensure you’re in a git repo.
-
-One scenario: very large diffs. Using `--compact-summary` and unified=3 limits output size, but if a commit touched 500 files, the diffstat could still be long.
-
-It might be worth limiting or truncating extremely large diffs before feeding to the LLM summary (perhaps summarizing the diffstat itself). This is more a potential enhancement than a bug; currently, a huge commit might risk hitting token limits for summarization. The code doesn’t explicitly guard against that beyond the diff settings.
-
-In practice, this might rarely be an issue, but it’s good to document.
-
-## Environment Variable Confusion:
-There are both config file `.env` loading and environment variables like `GIV_MODEL`, `GIV_API_URL`, etc. The code merges these by preferring CLI args, then env vars, then defaults.
-
-One thing to double-check: if the user sets `GIV_MODEL_MODE=none` (or uses `--model-mode none`), the script sets dry_run=true for generation but still goes through summarization unless explicitly handled.
-
-Actually, in generate_response, if mode is "none", it doesn’t have a case and will default to local which tries Ollama (not desired). There is a check setting model_mode="none" will skip model usage and treat it effectively as dry-run.
-
-This area could be refactored to make the logic clearer: e.g. a single flag that indicates “no model calls” which both summarization and final generation honor. As is, it does seem to work: model_mode none triggers warnings and then they set dry_run=true which causes generate_from_prompt to just print the prompt.
-
-It might be cleaner if summarize_target also respected dry-run and skipped calling the model (currently if you do `--model-mode none`, it might still run summarize_commit which calls generate_response; though since they set model_mode="none" and pass that in, generate_response would return empty immediately or error).
-
-This flow is a bit convoluted. A refactor could short-circuit summarization when no model is to be used – instead, just concatenate raw commit messages or something.
-
-This ties into perhaps providing a mode where the tool generates a draft prompt for manual use (which it does via `--dry-run` or `--prompt-file` in document). Not a critical bug, but worth reviewing.
-
-
 
 ## Pending