From 26c8cb246159ab4f476d4926e9f1902d7b7e237d Mon Sep 17 00:00:00 2001
From: Stuart Axelbrooke <stuart@axelbrooke.com>
Date: Thu, 17 Jul 2025 23:28:51 -0700
Subject: [PATCH] Add cargo.toml generator so intellij isn't useless for rust

---
 .gitignore                               |   1 +
 tools/build_rules/generate_cargo_toml.py | 275 +++++++++++++++++++++++
 2 files changed, 276 insertions(+)
 create mode 100644 tools/build_rules/generate_cargo_toml.py

diff --git a/.gitignore b/.gitignore
index fb697d3..a06ef4e 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,3 +9,4 @@ examples/podcast_reviews/data
 .venv
 node_modules
 **/node_modules
+Cargo.toml
diff --git a/tools/build_rules/generate_cargo_toml.py b/tools/build_rules/generate_cargo_toml.py
new file mode 100644
index 0000000..61ed75c
--- /dev/null
+++ b/tools/build_rules/generate_cargo_toml.py
@@ -0,0 +1,275 @@
+#!/usr/bin/env python3
+"""
+Generate a Cargo.toml file from MODULE.bazel crate specifications.
+This is purely for IDE support - Bazel still handles the actual builds.
+"""
+
+import re
+import sys
+from pathlib import Path
+import fnmatch
+
+def load_gitignore_patterns():
+    """Load .gitignore patterns if available."""
+    gitignore_file = Path(".gitignore")
+    patterns = [
+        # Default patterns to ignore
+        "target/", ".git/", "bazel-*", "*.pyc", "__pycache__/",
+        "node_modules/", ".DS_Store"
+    ]
+
+    if gitignore_file.exists():
+        content = gitignore_file.read_text()
+        for line in content.splitlines():
+            line = line.strip()
+            if line and not line.startswith('#'):
+                patterns.append(line)
+
+    return patterns
+
+def should_ignore(path, ignore_patterns):
+    """Check if a path should be ignored based on gitignore-style patterns."""
+    path_str = str(path)
+    for pattern in ignore_patterns:
+        if fnmatch.fnmatch(path_str, pattern) or fnmatch.fnmatch(path.name, pattern):
+            return True
+        # Handle directory patterns
+        if pattern.endswith('/') and pattern[:-1] in path.parts:
+            return True
+    return False
+
+def find_rust_sources(ignore_patterns):
+    """Find Rust source files, respecting gitignore patterns."""
+    rust_files = []
+
+    for rs_file in Path('.').rglob('*.rs'):
+        if not should_ignore(rs_file, ignore_patterns):
+            rust_files.append(rs_file)
+
+    return rust_files
+
+def detect_project_structure(rust_files):
+    """Detect if this is a library, binary, or mixed project."""
+    structure = {
+        'lib': None,
+        'bins': [],
+        'examples': [],
+        'tests': []
+    }
+
+    for rs_file in rust_files:
+        parts = rs_file.parts
+
+        # Check for lib.rs
+        if rs_file.name == 'lib.rs':
+            structure['lib'] = rs_file
+
+        # Check for main.rs (could be src/main.rs or in subdirs)
+        elif rs_file.name == 'main.rs':
+            # Determine the binary name from the path
+            if len(parts) >= 2 and parts[-2] == 'src':
+                # src/main.rs - default binary
+                structure['bins'].append(('main', rs_file))
+            elif len(parts) >= 3 and parts[-3] == 'src' and parts[-2] == 'bin':
+                # src/bin/name/main.rs or similar
+                bin_name = parts[-2] if parts[-2] != 'bin' else rs_file.stem
+                structure['bins'].append((bin_name, rs_file))
+            else:
+                # Other main.rs files, use directory name
+                bin_name = parts[-2] if len(parts) > 1 else 'main'
+                structure['bins'].append((bin_name, rs_file))
+
+        # Check for other binaries (src/bin/*.rs)
+        elif len(parts) >= 3 and parts[-2] == 'bin' and parts[-3] == 'src':
+            bin_name = rs_file.stem
+            structure['bins'].append((bin_name, rs_file))
+
+        # Check for examples
+        elif 'examples' in parts:
+            structure['examples'].append(rs_file)
+
+        # Check for tests
+        elif 'tests' in parts or rs_file.name.startswith('test_'):
+            structure['tests'].append(rs_file)
+
+    return structure
+
+def parse_crate_specs(module_content):
+    """Extract crate specifications from MODULE.bazel content."""
+    crates = {}
+
+    # Find all crate.spec() calls
+    spec_pattern = r'crate\.spec\(\s*(.*?)\s*\)'
+    specs = re.findall(spec_pattern, module_content, re.DOTALL)
+
+    for spec in specs:
+        # Parse the spec parameters
+        package_match = re.search(r'package\s*=\s*"([^"]+)"', spec)
+        version_match = re.search(r'version\s*=\s*"([^"]+)"', spec)
+        features_match = re.search(r'features\s*=\s*\[(.*?)\]', spec, re.DOTALL)
+        default_features_match = re.search(r'default_features\s*=\s*False', spec)
+
+        if package_match and version_match:
+            package = package_match.group(1)
+            version = version_match.group(1)
+
+            crate_info = {"version": version}
+
+            # Handle features
+            if features_match:
+                features_str = features_match.group(1)
+                features = [f.strip().strip('"') for f in features_str.split(',') if f.strip()]
+                if features:
+                    crate_info["features"] = features
+
+            # Handle default-features = false
+            if default_features_match:
+                crate_info["default-features"] = False
+
+            crates[package] = crate_info
+
+    return crates
+    """Extract crate specifications from MODULE.bazel content."""
+    crates = {}
+
+    # Find all crate.spec() calls
+    spec_pattern = r'crate\.spec\(\s*(.*?)\s*\)'
+    specs = re.findall(spec_pattern, module_content, re.DOTALL)
+
+    for spec in specs:
+        # Parse the spec parameters
+        package_match = re.search(r'package\s*=\s*"([^"]+)"', spec)
+        version_match = re.search(r'version\s*=\s*"([^"]+)"', spec)
+        features_match = re.search(r'features\s*=\s*\[(.*?)\]', spec, re.DOTALL)
+        default_features_match = re.search(r'default_features\s*=\s*False', spec)
+
+        if package_match and version_match:
+            package = package_match.group(1)
+            version = version_match.group(1)
+
+            crate_info = {"version": version}
+
+            # Handle features
+            if features_match:
+                features_str = features_match.group(1)
+                features = [f.strip().strip('"') for f in features_str.split(',') if f.strip()]
+                if features:
+                    crate_info["features"] = features
+
+            # Handle default-features = false
+            if default_features_match:
+                crate_info["default-features"] = False
+
+            crates[package] = crate_info
+
+    return crates
+
+def generate_cargo_toml(crates, structure, project_name="databuild"):
+    """Generate Cargo.toml content from parsed crates and project structure."""
+    lines = [
+        f'[package]',
+        f'name = "{project_name}"',
+        f'version = "0.1.0"',
+        f'edition = "2021"',
+        f'',
+        f'# Generated from MODULE.bazel for IDE support only',
+        f'# Actual dependencies are managed by Bazel',
+        f''
+    ]
+
+    # Add library section if lib.rs exists
+    if structure['lib']:
+        lines.extend([
+            f'[lib]',
+            f'path = "{structure["lib"]}"',
+            f''
+        ])
+
+    # Add binary sections
+    for bin_name, bin_path in structure['bins']:
+        lines.extend([
+            f'[[bin]]',
+            f'name = "{bin_name}"',
+            f'path = "{bin_path}"',
+            f''
+        ])
+
+    # Add example sections
+    for example_path in structure['examples']:
+        example_name = example_path.stem
+        lines.extend([
+            f'[[example]]',
+            f'name = "{example_name}"',
+            f'path = "{example_path}"',
+            f''
+        ])
+
+    lines.append('[dependencies]')
+
+    # Simple dependencies first
+    for package, info in sorted(crates.items()):
+        if not isinstance(info, dict):
+            lines.append(f'{package} = "{info}"')
+
+    # Complex dependencies as tables
+    for package, info in sorted(crates.items()):
+        if isinstance(info, dict):
+            lines.append(f'')
+            lines.append(f'[dependencies.{package}]')
+            lines.append(f'version = "{info["version"]}"')
+
+            if "features" in info:
+                features_str = ', '.join(f'"{f}"' for f in info["features"])
+                lines.append(f'features = [{features_str}]')
+
+            if "default-features" in info:
+                lines.append(f'default-features = {str(info["default-features"]).lower()}')
+
+    return '\n'.join(lines)
+
+def main():
+    module_file = Path("MODULE.bazel")
+    if not module_file.exists():
+        print("MODULE.bazel not found in current directory")
+        sys.exit(1)
+
+    # Load gitignore patterns
+    ignore_patterns = load_gitignore_patterns()
+
+    # Find Rust source files
+    rust_files = find_rust_sources(ignore_patterns)
+    if not rust_files:
+        print("No Rust source files found")
+        sys.exit(1)
+
+    print(f"Found {len(rust_files)} Rust source files")
+
+    # Detect project structure
+    structure = detect_project_structure(rust_files)
+
+    # Parse MODULE.bazel for dependencies
+    content = module_file.read_text()
+    crates = parse_crate_specs(content)
+
+    if not crates:
+        print("No crate specifications found in MODULE.bazel")
+        sys.exit(1)
+
+    # Generate Cargo.toml
+    cargo_toml = generate_cargo_toml(crates, structure)
+
+    output_file = Path("Cargo.toml")
+    output_file.write_text(cargo_toml)
+
+    print(f"Generated Cargo.toml with:")
+    print(f"  - {len(crates)} dependencies")
+    if structure['lib']:
+        print(f"  - Library: {structure['lib']}")
+    if structure['bins']:
+        print(f"  - {len(structure['bins'])} binaries: {[name for name, _ in structure['bins']]}")
+    if structure['examples']:
+        print(f"  - {len(structure['examples'])} examples")
+    print("This file is for IDE support only - Bazel manages the actual build")
+
+if __name__ == "__main__":
+    main()
\ No newline at end of file