diff --git a/haiku_rag_slim/haiku/rag/converters/text_utils.py b/haiku_rag_slim/haiku/rag/converters/text_utils.py index 64994e31..eb521b50 100644 --- a/haiku_rag_slim/haiku/rag/converters/text_utils.py +++ b/haiku_rag_slim/haiku/rag/converters/text_utils.py @@ -18,57 +18,129 @@ class TextFileHandler: # Plain text extensions that we'll read directly text_extensions: ClassVar[list[str]] = [ ".astro", + ".bash", ".c", + ".clj", + ".cljs", ".cpp", + ".cs", ".css", + ".dart", + ".elm", + ".ex", + ".exs", + ".fs", + ".fsx", ".go", + ".gql", + ".graphql", + ".groovy", ".h", + ".hcl", ".hpp", + ".hs", ".java", + ".jl", ".js", ".json", ".kt", + ".less", + ".lua", ".mdx", ".mjs", + ".ml", + ".mli", + ".nim", + ".nix", ".php", + ".pl", + ".pm", + ".proto", + ".ps1", ".py", + ".r", ".rb", ".rs", + ".sass", + ".scala", + ".scss", + ".sh", + ".sql", ".svelte", ".swift", + ".tf", + ".toml", ".ts", ".tsx", ".txt", ".vue", + ".xml", ".yaml", ".yml", + ".zig", ] # Code file extensions with their markdown language identifiers code_markdown_identifier: ClassVar[dict[str, str]] = { ".astro": "astro", + ".bash": "bash", ".c": "c", + ".clj": "clojure", + ".cljs": "clojure", ".cpp": "cpp", + ".cs": "csharp", ".css": "css", + ".dart": "dart", + ".elm": "elm", + ".ex": "elixir", + ".exs": "elixir", + ".fs": "fsharp", + ".fsx": "fsharp", ".go": "go", + ".gql": "graphql", + ".graphql": "graphql", + ".groovy": "groovy", ".h": "c", + ".hcl": "hcl", ".hpp": "cpp", + ".hs": "haskell", ".java": "java", + ".jl": "julia", ".js": "javascript", ".json": "json", ".kt": "kotlin", + ".less": "less", + ".lua": "lua", ".mjs": "javascript", + ".ml": "ocaml", + ".mli": "ocaml", + ".nim": "nim", + ".nix": "nix", ".php": "php", + ".pl": "perl", + ".pm": "perl", + ".proto": "protobuf", + ".ps1": "powershell", ".py": "python", + ".r": "r", ".rb": "ruby", ".rs": "rust", + ".sass": "sass", + ".scala": "scala", + ".scss": "scss", + ".sh": "bash", + ".sql": "sql", ".svelte": "svelte", ".swift": "swift", + ".tf": "hcl", + ".toml": "toml", ".ts": "typescript", ".tsx": "tsx", ".vue": "vue", + ".xml": "xml", ".yaml": "yaml", ".yml": "yaml", + ".zig": "zig", } @staticmethod