Implemented a check for larger base64 encoded blobs

Author: VitjanZ

file_utils.py

@@ -8,8 +8,9 @@
 
 import plain_spec
 from plain2code_console import console
-from plain2code_exceptions import UnsupportedResourceType
+from plain2code_exceptions import UnsupportedBase64Content, UnsupportedResourceType
 from plain2code_nodes import Plain2CodeIncludeTag, Plain2CodeLoaderMixin
+from plain2code_utils import find_large_base64_blob
 from plain_modules import CODEPLAIN_MEMORY_SUBFOLDER, CODEPLAIN_METADATA_FOLDER
 
 BINARY_FILE_EXTENSIONS = [".pyc"]
@@ -196,15 +197,26 @@ def load_linked_resources(template_dirs: list[str], resources_list, module_name:
             )
 
         if content is None:
-            raise FileNotFoundError(f"""
+            raise FileNotFoundError(
+                f"""
                 Resource file {file_name} not found. Resource files are searched in the following order (highest to lowest precedence):
 
                 1. The directory containing your .plain file
                 2. The directory specified by --template-dir (if provided)
                 3. The built-in 'standard_template_library' directory
 
                 Please ensure that the resource exists in one of these locations, or specify the correct --template-dir if using custom templates.
-                """)
+                """
+            )
+
+        blob = find_large_base64_blob(content)
+        if blob is not None:
+            raise UnsupportedBase64Content(
+                f"Referenced resource '{file_name}' in module '{module_name}' contains a large "
+                f"base64-encoded blob ({len(blob)} characters), such as an embedded image. Inline "
+                "base64 data is not supported. Remove the base64 data "
+                "from the resource. If necessary, provide the binary file path in the specification."
+            )
 
         linked_resources[file_name] = content
 

plain2code.py

@@ -37,6 +37,7 @@
     PlainSyntaxError,
     RenderCancelledError,
     RenderingCreditBalanceTooLow,
+    UnsupportedBase64Content,
     UnsupportedResourceType,
 )
 from plain2code_logger import (
@@ -74,6 +75,7 @@
     NetworkConnectionError,
     ModuleDoesNotExistError,
     UnsupportedResourceType,
+    UnsupportedBase64Content,
     GitNotInstalledError,
     SystemExit,
 )

plain2code_exceptions.py

@@ -29,6 +29,11 @@ class PlainSyntaxError(Exception):
     pass
 
 
+class UnsupportedBase64Content(Exception):
+
+    pass
+
+
 class InternalClientError(Exception):
     pass
 

plain2code_utils.py

@@ -1,3 +1,21 @@
+import re
+from typing import Optional
+
+#
+MIN_BASE64_BLOB_LENGTH = 8192
+
+# Matches a long contiguous base64 / base64url run, optionally preceded by a data: URI header.
+_BASE64_BLOB_PATTERN = re.compile(
+    r"(?:data:[\w.+-]+/[\w.+-]+;base64,)?[A-Za-z0-9+/_-]{%d,}={0,2}" % MIN_BASE64_BLOB_LENGTH
+)
+
+
+def find_large_base64_blob(text: str) -> Optional[str]:
+    """Return the first contiguous base64 blob at or above the threshold, or None."""
+    match = _BASE64_BLOB_PATTERN.search(text)
+    return match.group(0) if match else None
+
+
 def format_duration_hms(total_seconds: int) -> str:
     """Format a duration in seconds as hours, minutes, and seconds (e.g. ``1h 2m 3.45s``, ``45.67s``)."""
     if total_seconds < 0:

plain_file.py

@@ -17,8 +17,9 @@
 import concept_utils
 import file_utils
 import plain_spec
-from plain2code_exceptions import ModuleDoesNotExistError, PlainSyntaxError
+from plain2code_exceptions import ModuleDoesNotExistError, PlainSyntaxError, UnsupportedBase64Content
 from plain2code_nodes import Plain2CodeIncludeTag, Plain2CodeLoaderMixin
+from plain2code_utils import find_large_base64_blob
 
 RESOURCE_MARKER = "[resource]"
 
@@ -528,6 +529,16 @@ def read_module_plain_source(module_name: str, template_dirs: list[str]) -> str:
     plain_source_text = file_utils.open_from(template_dirs, module_name + PLAIN_SOURCE_FILE_EXTENSION)
     if plain_source_text is None:
         raise ModuleDoesNotExistError(f"Module does not exist ({module_name}).")
+
+    blob = find_large_base64_blob(plain_source_text)
+    if blob is not None:
+        raise UnsupportedBase64Content(
+            f"Module '{module_name}' contains a base64-encoded blob ({len(blob)} characters) "
+            "inlined in the specification. This is not supported."
+            "Remove the base64 data from the .plain file or if necessary,"
+            "include the binary file path in the specification."
+        )
+
     return plain_source_text
 
 

tests/data/sample_base64_image.txt

@@ -4,7 +4,8 @@
 import pytest
 
 from file_utils import load_linked_resources
-from plain2code_exceptions import UnsupportedResourceType
+from plain2code_exceptions import UnsupportedBase64Content, UnsupportedResourceType
+from plain2code_utils import MIN_BASE64_BLOB_LENGTH
 
 
 @pytest.fixture
@@ -39,3 +40,42 @@ def test_load_linked_resources_binary_file_raises_unsupported_resource_type(temp
 def test_load_linked_resources_missing_file_raises_file_not_found(template_dir):
     with pytest.raises(FileNotFoundError):
         load_linked_resources([template_dir], [{"text": "Missing", "target": "missing.md"}], "my_thing")
+
+
+def test_load_linked_resources_base64_blob_raises(template_dir):
+    file_path = os.path.join(template_dir, "request.txt")
+    with open(file_path, "w") as f:
+        f.write("curl -d 'selfie_image=" + "A" * (MIN_BASE64_BLOB_LENGTH + 100) + "'")
+
+    with pytest.raises(UnsupportedBase64Content) as exc_info:
+        load_linked_resources([template_dir], [{"text": "Request", "target": "request.txt"}], "my_thing")
+
+    assert "request.txt" in str(exc_info.value)
+    assert "my_thing" in str(exc_info.value)
+
+
+def test_load_linked_resources_small_base64_allowed(template_dir):
+    file_path = os.path.join(template_dir, "token.txt")
+    with open(file_path, "w") as f:
+        f.write("token=" + "A" * (MIN_BASE64_BLOB_LENGTH - 100))
+
+    result = load_linked_resources([template_dir], [{"text": "Token", "target": "token.txt"}], "my_thing")
+
+    assert "token.txt" in result
+
+
+def test_load_linked_resources_real_base64_image_raises(template_dir):
+    # Real base64-encoded JPEG that made Gemini 3 fail, inlined into a curl example resource.
+    sample_path = os.path.join(os.path.dirname(__file__), "data", "sample_base64_image.txt")
+    with open(sample_path) as f:
+        blob = f.read()
+
+    file_path = os.path.join(template_dir, "face_match_request.txt")
+    with open(file_path, "w") as f:
+        f.write(f"curl -d 'selfie_image={blob}' https://example.com/face-match")
+
+    with pytest.raises(UnsupportedBase64Content) as exc_info:
+        load_linked_resources([template_dir], [{"text": "Request", "target": "face_match_request.txt"}], "face_match")
+
+    assert "face_match_request.txt" in str(exc_info.value)
+    assert str(len(blob)) in str(exc_info.value)

tests/test_file_utils.py

@@ -65,15 +65,17 @@ def test_single_file_change(temp_repo):
     result = diff(temp_repo, "1.1")
 
     assert "test.txt" in result
-    expected_diff = dedent("""
+    expected_diff = dedent(
+        """
         --- a/test.txt
         +++ b/test.txt
         @@ -1,3 +1,3 @@
         -initial content
         +modified content
          line2
          line3
-    """).strip()
+    """
+    ).strip()
     assert result["test.txt"] == expected_diff
 
 
@@ -98,27 +100,31 @@ def test_multiple_file_changes(temp_repo):
 
     # Check first file
     assert "test.txt" in result
-    expected_diff1 = dedent("""
+    expected_diff1 = dedent(
+        """
         --- a/test.txt
         +++ b/test.txt
         @@ -1,3 +1,2 @@
         -initial content
         +file1 modified
          line2
         -line3
-    """).strip()
+    """
+    ).strip()
     assert result["test.txt"] == expected_diff1
 
     # Check second file
     assert "file2.txt" in result
-    expected_diff2 = dedent("""
+    expected_diff2 = dedent(
+        """
         --- /dev/null
         +++ b/file2.txt
         @@ -0,0 +1,2 @@
         +file2 modified
         +line2
         \\ No newline at end of file
-    """).strip()
+    """
+    ).strip()
     assert result["file2.txt"] == expected_diff2
 
 
@@ -153,38 +159,44 @@ def test_multiple_commits_diff(temp_repo):
 
     # Check first file
     assert "test.txt" in result
-    expected_diff1 = dedent("""
+    expected_diff1 = dedent(
+        """
         --- a/test.txt
         +++ b/test.txt
         @@ -1,3 +1,2 @@
         -initial content
         +file1 frid1.2 refactored version
          line2
         -line3
-    """).strip()
+    """
+    ).strip()
     assert result["test.txt"] == expected_diff1
 
     # Check second file
     assert "file2.txt" in result
-    expected_diff2 = dedent("""
+    expected_diff2 = dedent(
+        """
         --- a/file2.txt
         +++ b/file2.txt
         @@ -1,2 +1,2 @@
         -file2 frid1.1 refactored version
         +file2 frid1.2 version
          line2
-    """).strip()
+    """
+    ).strip()
     assert result["file2.txt"] == expected_diff2
 
     # Check third file
     assert "file3.txt" in result
-    expected_diff3 = dedent("""
+    expected_diff3 = dedent(
+        """
         --- /dev/null
         +++ b/file3.txt
         @@ -0,0 +1,2 @@
         +file3 frid1.2 new file
         +line2
-    """).strip()
+    """
+    ).strip()
     assert result["file3.txt"] == expected_diff3
 
 
@@ -203,23 +215,27 @@ def test_diff_without_previous_frid_and_no_base_folder(empty_repo):
     result = diff(empty_repo)
 
     assert "new.txt" in result
-    expected_diff = dedent("""
+    expected_diff = dedent(
+        """
         --- /dev/null
         +++ b/new.txt
         @@ -0,0 +1,2 @@
         +new file content
         +line2
-    """).strip()
+    """
+    ).strip()
     assert result["new.txt"] == expected_diff
 
     assert "new2.txt" in result
-    expected_diff2 = dedent("""
+    expected_diff2 = dedent(
+        """
         --- /dev/null
         +++ b/new2.txt
         @@ -0,0 +1,2 @@
         +new file content
         +line2
-    """).strip()
+    """
+    ).strip()
     assert result["new2.txt"] == expected_diff2
 
 
@@ -237,14 +253,16 @@ def test_diff_without_previous_frid_and_base_folder(temp_repo):
     result = diff(temp_repo)
 
     assert "new.txt" in result
-    expected_diff = dedent("""
+    expected_diff = dedent(
+        """
         --- a/new.txt
         +++ b/new.txt
         @@ -1,2 +1,2 @@
         -base folder content
         +updated base folder content
          line2
-    """).strip()
+    """
+    ).strip()
     assert result["new.txt"] == expected_diff
 
 
@@ -260,13 +278,15 @@ def test_new_file(temp_repo):
     result = diff(temp_repo, "1.1")
 
     assert "new.txt" in result
-    expected_diff = dedent("""
+    expected_diff = dedent(
+        """
         --- /dev/null
         +++ b/new.txt
         @@ -0,0 +1,2 @@
         +new file content
         +line2
-    """).strip()
+    """
+    ).strip()
     assert result["new.txt"] == expected_diff
 
 
@@ -282,14 +302,16 @@ def test_deleted_file(temp_repo):
     result = diff(temp_repo, "1.1")
 
     assert "test.txt" in result
-    expected_diff = dedent("""
+    expected_diff = dedent(
+        """
         --- a/test.txt
         +++ /dev/null
         @@ -1,3 +0,0 @@
         -initial content
         -line2
         -line3
-    """).strip()
+    """
+    ).strip()
     assert result["test.txt"] == expected_diff