Simplify context script with clearer options and focused design

vgrichina · vgrichina · commit e9e8446b56b8 · 2025-03-19T04:46:26.000-07:00
diff --git a/context b/context
@@ -1,57 +1,52 @@
 #!/bin/bash
 
-# context - Code Context Generator
+# context - Simplified Code Context Generator
 # Generates contextual information from a codebase to send to an LLM
 
 # Default values
-FILES=()
-EXCLUDE=""
+INCLUDE=()
+EXCLUDE=()
 MAX_SIZE="500KB"
-DEPTH=1
 INCLUDE_GIT=false
-GIT_DEPTH=3
-SUMMARY=false
 SHOW_FILE_SIZES=false
-TRUNCATE_LARGE=""
-INCLUDE_LS_FILES=true
-LS_FILES_LIMIT=100
+SHOW_LS_FILES=true
 
 # Parse arguments
 while [[ "$#" -gt 0 ]]; do
     case $1 in
-        --files=*) FILES+=("${1#*=}"); shift ;;
-        --exclude=*) EXCLUDE="${1#*=}"; shift ;;
+        --include=*) INCLUDE+=("${1#*=}"); shift ;;
+        --exclude=*) EXCLUDE+=("${1#*=}"); shift ;;
         --max-size=*) MAX_SIZE="${1#*=}"; shift ;;
-        --depth=*) DEPTH="${1#*=}"; shift ;;
-        --include-git) INCLUDE_GIT=true; shift ;;
-        --git-depth=*) GIT_DEPTH="${1#*=}"; shift ;;
-        --summary) SUMMARY=true; shift ;;
-        --show-file-sizes) SHOW_FILE_SIZES=true; shift ;;
-        --truncate-large=*) TRUNCATE_LARGE="${1#*=}"; shift ;;
-        --ls-files) INCLUDE_LS_FILES=true; shift ;;
-        --no-ls-files) INCLUDE_LS_FILES=false; shift ;;
-        --ls-files-limit=*) LS_FILES_LIMIT="${1#*=}"; shift ;;
+        --git) INCLUDE_GIT=true; shift ;;
+        --show-sizes) SHOW_FILE_SIZES=true; shift ;;
+        --no-ls-files) SHOW_LS_FILES=false; shift ;;
         --help|-h)
-            echo "Usage: ./context [options] [file1 file2 ...]"
+            echo "Usage: ./context [options] [include-pattern1 include-pattern2 ...]"
             echo ""
             echo "Options:"
-            echo "  --files=<pattern>        File pattern to include (e.g., \"src/*.js\")"
+            echo "  --include=<pattern>      File pattern to include (e.g., \"src/*.js\")"
             echo "  --exclude=<pattern>      File pattern to exclude (e.g., \"node_modules/**\")"
             echo "  --max-size=<size>        Maximum context size in KB/MB (e.g., \"500KB\")"
-            echo "  --depth=<num>            Dependency traversal depth (default: 1)"
-            echo "  --include-git            Include git information (recent commits, authors)"
-            echo "  --git-depth=<num>        Number of recent commits to include (default: 3)"
-            echo "  --summary                Include short summary of each file"
-            echo "  --show-file-sizes        Include file sizes in output"
-            echo "  --truncate-large=<size>  Truncate files larger than specified size (e.g., \"50KB\")"
-            echo "  --ls-files               Include git ls-files output (default: true)"
-            echo "  --no-ls-files            Don't include git ls-files output"
-            echo "  --ls-files-limit=<num>   Limit the number of files shown in ls-files (default: 100)"
+            echo "  --git                    Include basic git information (recent commits, branch)"
+            echo "  --show-sizes             Include file sizes in output"
+            echo "  --no-ls-files            Don't include git ls-files output in repository map"
             echo "  --help, -h               Show this help message"
+            echo ""
+            echo "Pattern matching:"
+            echo "  Patterns use bash 'find' command syntax with -path flag"
+            echo "  Examples:"
+            echo "    \"*.js\"               - All JavaScript files in current directory"
+            echo "    \"src/*.js\"           - All JavaScript files in src directory" 
+            echo "    \"src/**/*.js\"        - All JavaScript files in src and subdirectories"
+            echo "    \"!test/*\"            - Exclude all files in test directory"
+            echo ""
+            echo "Examples:"
+            echo "  ./context --include=\"src/*.js\" --exclude=\"*.test.js\""
+            echo "  ./context \"src/*.js\" \"*.md\" --max-size=1MB"
             exit 0
             ;;
         --*) echo "Unknown parameter: $1"; exit 1 ;;
-        *) FILES+=("$1"); shift ;;
+        *) INCLUDE+=("$1"); shift ;;
     esac
 done
 
@@ -70,30 +65,28 @@ convert_to_bytes() {
 }
 
 MAX_SIZE_BYTES=$(convert_to_bytes "$MAX_SIZE")
-TRUNCATE_SIZE_BYTES=0
-if [ -n "$TRUNCATE_LARGE" ]; then
-    TRUNCATE_SIZE_BYTES=$(convert_to_bytes "$TRUNCATE_LARGE")
-fi
 
 # Get all matching files
 find_files() {
     local all_files=""
     
-    for file_pattern in "${FILES[@]}"; do
-        if [ -f "$file_pattern" ]; then
-            all_files="$all_files"$'\n'"$file_pattern"
+    for pattern in "${INCLUDE[@]}"; do
+        if [ -f "$pattern" ]; then
+            all_files="$all_files"$'\n'"$pattern"
         else
-            local find_cmd="find . -type f -path \"$file_pattern\""
-            if [ -n "$EXCLUDE" ]; then
-                find_cmd="$find_cmd -not -path \"$EXCLUDE\""
-            fi
+            local find_cmd="find . -type f -path \"$pattern\""
             local found_files=$(eval $find_cmd)
             if [ -n "$found_files" ]; then
                 all_files="$all_files"$'\n'"$found_files"
             fi
         fi
     done
     
+    # Process exclusions
+    for exclude_pattern in "${EXCLUDE[@]}"; do
+        all_files=$(echo "$all_files" | grep -v "$exclude_pattern")
+    done
+    
     echo "$all_files" | grep -v "^$" | sort | uniq
 }
 
@@ -119,11 +112,35 @@ human_readable_size() {
 # Start Markdown output
 echo "# Code Context"
 echo ""
-if [ ${#FILES[@]} -gt 0 ]; then
-    echo "## Files"
+
+# Include git information if requested
+if [ "$INCLUDE_GIT" = true ] && command -v git >/dev/null 2>&1 && git rev-parse --is-inside-work-tree >/dev/null 2>&1; then
+    echo "## Repository Information"
+    echo ""
+    echo "Branch: $(git branch --show-current)"
+    echo ""
+    echo "Recent commits:"
+    echo ""
+    git log -n 3 --pretty=format:"* %h: %s (%an, %ar)" | while read line; do
+        echo "$line"
+    done
+    echo ""
     echo ""
 fi
 
+# Include git ls-files by default for repository map
+if [ "$SHOW_LS_FILES" = true ] && command -v git >/dev/null 2>&1 && git rev-parse --is-inside-work-tree >/dev/null 2>&1; then
+    echo "## Repository Map"
+    echo ""
+    git ls-files | sort | while read -r file; do
+        echo "- $file"
+    done
+    echo ""
+fi
+
+echo "## Files"
+echo ""
+
 # Process each file
 for file in $ALL_FILES; do
     if [ ! -f "$file" ]; then
@@ -134,92 +151,24 @@ for file in $ALL_FILES; do
     TOTAL_SIZE=$((TOTAL_SIZE + file_size))
     
     if [ $TOTAL_SIZE -gt $MAX_SIZE_BYTES ]; then
-        echo "Error: Total context size ($TOTAL_SIZE bytes) exceeds maximum allowed size ($(human_readable_size $MAX_SIZE_BYTES)). Reduce file scope or increase --max-size." >&2
+        echo "Error: Total context size exceeds maximum allowed size ($(human_readable_size $MAX_SIZE_BYTES)). Reduce file scope or increase --max-size." >&2
         exit 1
     fi
     
-    truncated=false
-    file_content=""
-    
-    if [ $TRUNCATE_SIZE_BYTES -gt 0 ] && [ $file_size -gt $TRUNCATE_SIZE_BYTES ]; then
-        file_content=$(head -c $TRUNCATE_SIZE_BYTES "$file")
-        truncated=true
-    else
-        file_content=$(cat "$file")
-    fi
-    
-    file_summary=""
-    if [ "$SUMMARY" = true ]; then
-        file_summary=$(head -n 20 "$file" | grep -E "^(//|#|/*) " | head -n 5 | sed 's/^[\/\#\* ]*//')
-    fi
+    file_content=$(cat "$file")
     
     if [ "$SHOW_FILE_SIZES" = true ]; then
-        echo "Size: $(human_readable_size $file_size)"
-        echo ""
-    fi
-    
-    if [ "$SUMMARY" = true ] && [ -n "$file_summary" ]; then
-        echo "Summary:"
-        echo "$file_summary"
-        echo ""
+        echo "### $file ($(human_readable_size $file_size))"
+    else
+        echo "### $file"
     fi
+    echo ""
     
-    # Use proper markdown code block with three backticks
-    echo "\`\`\` ${file##*.} $file"
+    # Use proper markdown code block with file extension for syntax highlighting
+    echo "\`\`\` ${file##*.}"
     echo "$file_content"
-    if [ "$truncated" = true ]; then
-        echo ""
-        echo "... [File truncated due to size limit] ..."
-    fi
     echo "\`\`\`"
     echo ""
 done
 
-# Include git information if requested
-if [ "$INCLUDE_GIT" = true ] && command -v git >/dev/null 2>&1 && git rev-parse --is-inside-work-tree >/dev/null 2>&1; then
-    echo "## Git Information"
-    echo ""
-    echo "### Recent Commits"
-    echo ""
-    git log -n $GIT_DEPTH --pretty=format:"* %h: %s (%an, %ar)" | while read line; do
-        echo "$line"
-    done
-    echo ""
-    echo "### Branch Information"
-    echo ""
-    echo "Current branch: $(git branch --show-current)"
-    echo ""
-fi
-
-# Include git ls-files output if requested
-if [ "$INCLUDE_LS_FILES" = true ] && command -v git >/dev/null 2>&1 && git rev-parse --is-inside-work-tree >/dev/null 2>&1; then
-    GIT_FILES=$(git ls-files | sort)
-    GIT_FILES_COUNT=$(echo "$GIT_FILES" | wc -l)
-    
-    if [ $GIT_FILES_COUNT -gt $LS_FILES_LIMIT ]; then
-        GIT_FILES=$(echo "$GIT_FILES" | head -n $LS_FILES_LIMIT)
-        GIT_FILES_TRUNCATED=true
-    else
-        GIT_FILES_TRUNCATED=false
-    fi
-    
-    echo "## Repository Files"
-    echo ""
-    echo "$GIT_FILES" | while read -r file; do
-        echo "- $file"
-    done
-    if [ "$GIT_FILES_TRUNCATED" = true ]; then
-        echo "- ... (and $(($GIT_FILES_COUNT - $LS_FILES_LIMIT)) more files)"
-    fi
-    echo ""
-fi
-
-# Include default prompt if available
-if [ -f "prompts/context_prompt.txt" ]; then
-    echo "## Instructions for LLM"
-    echo ""
-    cat "prompts/context_prompt.txt"
-    echo ""
-fi
-
 exit 0
diff --git a/docs/context.md b/docs/context.md
@@ -1,112 +1,17 @@
-# `context` - Code Context Generator
+# `context` - Simplified Code Context Generator
 
 ## Overview
 
-The `context` tool extracts relevant code and context from your codebase to send to an LLM (Large Language Model). It helps create focused, comprehensive snapshots of your codebase for more accurate AI assistance.
+The `context` tool extracts relevant code from your codebase to send to an LLM (Large Language Model). It creates focused snapshots of your code for more accurate AI assistance with a streamlined interface.
 
 ## Usage
 
 ```bash
-# Using patterns
-./context --files="src/components/*.js" --exclude="*.test.js" --max-size=300KB > context.txt
+# Using file patterns
+./context --include="src/*.js" --exclude="*.test.js" > context.md
 
-# Using direct file paths
-./context src/app.js src/utils.js README.md > context.txt
+# Using direct patterns as arguments
+./context "src/*.js" "README.md" > context.md
 
-# Including git information and custom depth
-./context --files="src/*.js" --include-git --git-depth=5 --depth=2 > context.txt
-```
-
-## Arguments
-
-| Argument | Description | Default |
-|----------|-------------|---------|
-| `--files=<pattern>` | File pattern to include (e.g., "src/*.js") | None |
-| Direct file arguments | Files to include (e.g., app.js README.md) | None |
-| `--exclude=<pattern>` | File pattern to exclude (e.g., "node_modules/**") | None |
-| `--max-size=<size>` | Maximum context size in KB/MB (e.g., "500KB") | "500KB" |
-| `--depth=<num>` | Dependency traversal depth | 1 |
-| `--include-git` | Include git information (recent commits, authors) | False |
-| `--git-depth=<num>` | Number of recent commits to include | 3 |
-| `--summary` | Include short summary of each file | False |
-| `--show-file-sizes` | Include file sizes in output | False |
-| `--truncate-large=<size>` | Truncate files larger than specified size (e.g., "50KB") | None |
-| `--verbose` | Show verbose output during processing | False |
-| `--ls-files` | Include git ls-files output | True |
-| `--no-ls-files` | Don't include git ls-files output | False |
-| `--ls-files-limit=<num>` | Limit the number of files shown in ls-files | 100 |
-| `--help`, `-h` | Show help message | - |
-
-## Examples
-
-### Basic Usage
-
-Extract all JS files in the src directory:
-
-```bash
-./context --files="src/**/*.js" > code_context.md
-```
-
-### Exclude Specific Files
-
-Include all JavaScript files but exclude test files:
-
-```bash
-./context --files="src/**/*.js" --exclude="**/*.test.js" > context.md
-```
-
-### Custom Size Limits
-
-Set maximum context size and truncate large files:
-
-```bash
-./context --files="src/*.js" --max-size=1MB --truncate-large=50KB > context.md
-```
-
-### Git Information
-
-Generate context with git history:
-
-```bash
-./context --files="src/utils/*.js" --include-git --git-depth=5 > utils_context.md
-```
-
-### File Summary and Sizes
-
-Include file summaries and size information:
-
-```bash
-./context --files="src/*.py" --summary --show-file-sizes > context.md
-```
-
-### Control Repository File Listing
-
-Limit the number of files shown in the repository listing:
-
-```bash
-./context --files="src/main.js" --ls-files-limit=50 > context.md
-```
-
-Or disable the repository file listing completely:
-
-```bash
-./context --files="src/main.js" --no-ls-files > context.md
-```
-
-## Output Format
-
-The tool generates markdown output with the following sections:
-
-1. **Code Context** - The main header
-2. **Files** - The content of each matched file
-3. **Git Information** (if requested) - Recent commits and branch information
-4. **Repository Files** - List of files in the git repository
-5. **Instructions for LLM** (if available) - Instructions from a prompt template
-
-## Tips
-
-- Keep the context focused and relevant for the task at hand
-- Use `--exclude` to filter out test files, build artifacts, and other non-essential files
-- The `--summary` flag helps provide high-level context about each file
-- If you hit token limits with your LLM, use `--truncate-large` to limit file sizes
-- For large repositories, adjust `--ls-files-limit` to show only the most relevant files
+# Including git information
+./context --include="src/*.js" --git > context.md
