Update Python tools

python-tools/app/discovery.py

@@ -11,7 +11,7 @@ from typing import Any
 
 
 FIELD_NAME = re.compile(r"[a-z][a-z0-9_]*\Z")
-FIELD_TYPES = frozenset({"text", "integer", "date", "datetime", "choice", "multi_choice"})
+FIELD_TYPES = frozenset({"text", "integer", "date", "datetime", "choice", "multi_choice", "file"})
 
 
 def display_name(value: str) -> str:
@@ -28,6 +28,7 @@ class Task:
     description: str | None
     inputs: tuple["InputField", ...]
     wait_for_result: bool = False
+    download_artifacts: bool = False
 
 
 @dataclass(frozen=True)
@@ -58,6 +59,8 @@ class InputField:
     step: int | None
     pattern: str | None
     options: tuple[InputOption, ...]
+    accept: tuple[str, ...]
+    maximum_bytes: int | None
 
 
 @dataclass(frozen=True)
@@ -154,6 +157,7 @@ class TaskCatalog:
                     description=read_description(path),
                     inputs=declaration.inputs if declaration else (),
                     wait_for_result=declaration.wait_for_result if declaration else False,
+                    download_artifacts=declaration.download_artifacts if declaration else False,
                 )
             )
         return tasks
@@ -275,6 +279,7 @@ def load_input_manifest(
 class TaskDeclaration:
     inputs: tuple[InputField, ...]
     wait_for_result: bool
+    download_artifacts: bool
 
 
 def parse_task_declaration(
@@ -282,7 +287,11 @@ def parse_task_declaration(
     filename: str,
     declaration: Any,
 ) -> TaskDeclaration:
-    if not isinstance(declaration, dict) or not {"inputs"} <= set(declaration) or set(declaration) - {"inputs", "wait_for_result"}:
+    if (
+        not isinstance(declaration, dict)
+        or not {"inputs"} <= set(declaration)
+        or set(declaration) - {"inputs", "wait_for_result", "download_artifacts"}
+    ):
         raise ValueError(f"{manifest_path} task {filename} has an invalid declaration.")
     raw_inputs = declaration["inputs"]
     if not isinstance(raw_inputs, list):
@@ -294,7 +303,16 @@ def parse_task_declaration(
     wait_for_result = declaration.get("wait_for_result", False)
     if not isinstance(wait_for_result, bool):
         raise ValueError(f"{manifest_path} task {filename} wait_for_result must be true or false.")
-    return TaskDeclaration(inputs=fields, wait_for_result=wait_for_result)
+    download_artifacts = declaration.get("download_artifacts", False)
+    if not isinstance(download_artifacts, bool):
+        raise ValueError(f"{manifest_path} task {filename} download_artifacts must be true or false.")
+    if download_artifacts and not wait_for_result:
+        raise ValueError(f"{manifest_path} task {filename} download_artifacts requires wait_for_result.")
+    return TaskDeclaration(
+        inputs=fields,
+        wait_for_result=wait_for_result,
+        download_artifacts=download_artifacts,
+    )
 
 
 def parse_input_field(manifest_path: Path, filename: str, raw: Any) -> InputField:
@@ -302,7 +320,7 @@ def parse_input_field(manifest_path: Path, filename: str, raw: Any) -> InputFiel
         raise ValueError(f"{manifest_path} task {filename} has a non-object input field.")
     allowed = {
         "name", "label", "type", "required", "default", "minimum", "maximum",
-        "step", "pattern", "options",
+        "step", "pattern", "options", "accept", "maximum_bytes",
     }
     unknown = set(raw) - allowed
     if unknown:
@@ -344,6 +362,20 @@ def parse_input_field(manifest_path: Path, filename: str, raw: Any) -> InputFiel
         except re.error as error:
             raise ValueError(f"{manifest_path} task {filename} input {name} has an invalid pattern: {error}.") from error
 
+    accept = parse_file_accept(manifest_path, filename, name, field_type, raw.get("accept"))
+    maximum_bytes = raw.get("maximum_bytes")
+    if field_type == "file":
+        if pattern is not None or raw.get("default") is not None:
+            raise ValueError(f"{manifest_path} task {filename} file input {name} cannot define a pattern or default.")
+        if maximum_bytes is not None and (
+            not isinstance(maximum_bytes, int)
+            or isinstance(maximum_bytes, bool)
+            or maximum_bytes < 1
+        ):
+            raise ValueError(f"{manifest_path} task {filename} file input {name} maximum_bytes must be positive.")
+    elif maximum_bytes is not None:
+        raise ValueError(f"{manifest_path} task {filename} input {name} only file fields accept maximum_bytes.")
+
     options = parse_options(manifest_path, filename, name, field_type, raw.get("options"))
     default = parse_default(manifest_path, filename, name, field_type, options, raw.get("default"))
     validate_default(
@@ -368,9 +400,34 @@ def parse_input_field(manifest_path: Path, filename: str, raw: Any) -> InputFiel
         step=numeric_values["step"],
         pattern=pattern,
         options=options,
+        accept=accept,
+        maximum_bytes=maximum_bytes,
     )
 
 
+def parse_file_accept(
+    manifest_path: Path,
+    filename: str,
+    name: str,
+    field_type: str,
+    raw_accept: Any,
+) -> tuple[str, ...]:
+    if field_type != "file":
+        if raw_accept is not None:
+            raise ValueError(f"{manifest_path} task {filename} input {name} only file fields accept accept.")
+        return ()
+    if raw_accept is None:
+        return ()
+    if (
+        not isinstance(raw_accept, list)
+        or not raw_accept
+        or not all(isinstance(value, str) and re.fullmatch(r"\.[a-z0-9]{1,10}", value) for value in raw_accept)
+        or len(set(raw_accept)) != len(raw_accept)
+    ):
+        raise ValueError(f"{manifest_path} task {filename} file input {name} has invalid accept values.")
+    return tuple(raw_accept)
+
+
 def parse_options(
     manifest_path: Path,
     filename: str,