feat: enhance file upload process with content type detection and command building

api/core/sandbox/inspector/archive_source.py

@@ -10,6 +10,8 @@ from core.sandbox.inspector.base import SandboxFileSource
 from core.sandbox.inspector.script_utils import (
     build_detect_kind_command,
     build_list_command,
+    build_upload_command,
+    guess_content_type,
     parse_kind_output,
     parse_list_output,
 )
@@ -129,8 +131,9 @@ class SandboxFileArchiveSource(SandboxFileSource):
             sandbox_storage = SandboxFileService.get_storage()
             is_file = kind == "file"
             filename = (os.path.basename(path) or "file") if is_file else f"{export_name}.tar.gz"
-            export_key = SandboxFilePaths.export(self._tenant_id, self._app_id, self._sandbox_id, export_id, filename)
+            export_key = SandboxFilePaths.export(self._tenant_id, self._app_id, self._sandbox_id, export_id)
             upload_url = sandbox_storage.get_upload_url(export_key, self._EXPORT_EXPIRES_IN_SECONDS)
+            content_type = guess_content_type(filename)
 
             # Build pipeline: for directories, tar first then upload; for files, upload directly
             archive_temp = f"/tmp/{export_id}.tar.gz"
@@ -146,7 +149,7 @@ class SandboxFileArchiveSource(SandboxFileSource):
                         on=not is_file,
                     )
                     .add(
-                        ["curl", "-sf", "-X", "PUT", "-T", src_path, upload_url],
+                        build_upload_command(src_path, upload_url, content_type=content_type),
                         error_message="Failed to upload file",
                     )
                     .add(["rm", "-f", archive_temp], on=not is_file)

api/core/sandbox/inspector/runtime_source.py

@@ -9,6 +9,8 @@ from core.sandbox.inspector.base import SandboxFileSource
 from core.sandbox.inspector.script_utils import (
     build_detect_kind_command,
     build_list_command,
+    build_upload_command,
+    guess_content_type,
     parse_kind_output,
     parse_list_output,
 )
@@ -88,11 +90,11 @@ class SandboxFileRuntimeSource(SandboxFileSource):
             self._app_id,
             self._sandbox_id,
             export_id,
-            filename,
         )
 
         sandbox_storage = SandboxFileService.get_storage()
         upload_url = sandbox_storage.get_upload_url(export_key, self._EXPORT_EXPIRES_IN_SECONDS)
+        content_type = guess_content_type(filename)
 
         if kind == "dir":
             archive_path = f"/tmp/{export_id}.tar.gz"
@@ -104,7 +106,7 @@ class SandboxFileRuntimeSource(SandboxFileSource):
                         error_message="Failed to archive directory in sandbox",
                     )
                     .add(
-                        ["curl", "-s", "-f", "-X", "PUT", "-T", archive_path, upload_url],
+                        build_upload_command(archive_path, upload_url, content_type=content_type),
                         error_message="Failed to upload directory archive from sandbox",
                     )
                     .execute(timeout=self._UPLOAD_TIMEOUT_SECONDS, raise_on_error=True)
@@ -124,7 +126,7 @@ class SandboxFileRuntimeSource(SandboxFileSource):
                 (
                     pipeline(self._runtime)
                     .add(
-                        ["curl", "-s", "-f", "-X", "PUT", "-T", path, upload_url],
+                        build_upload_command(path, upload_url, content_type=content_type),
                         error_message="Failed to upload file from sandbox",
                     )
                     .execute(timeout=self._UPLOAD_TIMEOUT_SECONDS, raise_on_error=True)

api/core/sandbox/inspector/script_utils.py

@@ -3,6 +3,7 @@
 from __future__ import annotations
 
 import json
+import mimetypes
 from typing import TypedDict, cast
 
 _PYTHON_EXEC_CMD = 'if command -v python3 >/dev/null 2>&1; then py=python3; else py=python; fi; "$py" -c "$0" "$@"'
@@ -96,3 +97,22 @@ def parse_kind_output(stdout: bytes, *, not_found_message: str) -> str:
     if kind not in ("dir", "file"):
         raise ValueError(not_found_message)
     return kind
+
+
+def guess_content_type(filename: str) -> str | None:
+    content_type, _ = mimetypes.guess_type(filename, strict=False)
+    if content_type is None:
+        return None
+    if content_type.startswith("text/"):
+        return f"{content_type}; charset=utf-8"
+    if content_type == "application/json":
+        return "application/json; charset=utf-8"
+    return content_type
+
+
+def build_upload_command(src_path: str, upload_url: str, *, content_type: str | None) -> list[str]:
+    command = ["curl", "-s", "-f", "-X", "PUT", "-T", src_path]
+    if content_type:
+        command.extend(["-H", f"Content-Type: {content_type}"])
+    command.append(upload_url)
+    return command

api/core/sandbox/storage/sandbox_file_storage.py

@@ -11,9 +11,9 @@ class SandboxFilePaths:
     """Facade for generating sandbox file storage keys."""
 
     @staticmethod
-    def export(tenant_id: str, app_id: str, sandbox_id: str, export_id: str, filename: str) -> str:
+    def export(tenant_id: str, app_id: str, sandbox_id: str, export_id: str) -> str:
         """sandbox_files/{tenant}/{app}/{sandbox}/{export_id}/{filename}"""
-        return f"sandbox_files/{tenant_id}/{app_id}/{sandbox_id}/{export_id}/{filename}"
+        return f"sandbox_files/{tenant_id}/{app_id}/{sandbox_id}/{export_id}"
 
     @staticmethod
     def archive(tenant_id: str, app_id: str, sandbox_id: str) -> str: