feat(gemini-image): add Gemini image generation package#3
Conversation
Add new workspace package for image generation using Google's Gemini models (Nano Banana / Nano Banana Pro). Features: - Text-to-image generation with configurable resolution (1K/2K/4K) - Multiple aspect ratios (1:1, 3:4, 4:3, 9:16, 16:9) - Reference-based image editing and refinement - Multi-part story sequence generation with visual continuity - Draft-then-finalize workflow for 75% cost reduction - Thinking mode with intermediate image visualization - CLI tool (`gemini-image`) for command-line usage Package structure: - generator.py: Core generate_image(), generate_story_sequence(), finalize_draft() functions - models.py: Model configurations and type definitions - utils.py: Helper functions for API key, base64, file extensions - cli.py: Full-featured command-line interface Models supported: - flash: Gemini 2.5 Flash (fast generation) - pro: Gemini 3 Pro (4K, better text, thinking mode) [default] Based on handoff from /home/byron/dev/library/scripts/generate_image.py 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
WalkthroughIntroduces a new Changes
Sequence Diagram(s)sequenceDiagram
participant User
participant CLI
participant Generator
participant Utils
participant GeminiAPI as Gemini API
participant FileSystem as File System
User->>CLI: invoke generate_image()
CLI->>Utils: get_api_key()
Utils-->>CLI: API key
CLI->>Generator: generate_image(prompt, model, ...)
Generator->>Utils: load_image_as_base64(ref_image)
Utils-->>Generator: base64_data, mime_type
Generator->>GeminiAPI: generate_content(contents, config)
GeminiAPI-->>Generator: response with image parts
Generator->>Generator: iterate parts, extract thoughts & images
Generator->>Utils: decode_base64_image()
Utils-->>Generator: image bytes
Generator->>FileSystem: write image file
FileSystem-->>Generator: file path
Generator-->>CLI: return Path
CLI-->>User: output file path
sequenceDiagram
participant User
participant CLI
participant Generator
participant GeminiAPI as Gemini API
participant FileSystem as File System
User->>CLI: invoke generate_story_sequence(base_prompt, 3 parts)
CLI->>Generator: generate_story_sequence(...)
loop for each part (1 to 3)
Generator->>Generator: construct part-specific prompt
Generator->>GeminiAPI: generate_content(prompt, prev_image as ref)
GeminiAPI-->>Generator: response with image
Generator->>FileSystem: write part_N.png
FileSystem-->>Generator: file path
Generator->>Generator: use output as reference for next part
end
Generator-->>CLI: return list of Paths
CLI-->>User: output paths for all parts
sequenceDiagram
participant User
participant CLI
participant Generator
participant FileSystem as File System
participant GeminiAPI as Gemini API
User->>CLI: invoke finalize_draft(draft.png, prompt)
CLI->>Generator: finalize_draft(draft_path, ...)
Generator->>FileSystem: validate draft exists
FileSystem-->>Generator: draft file
Generator->>Generator: load draft as reference
Generator->>GeminiAPI: generate_content(prompt, draft_ref, 2K resolution)
GeminiAPI-->>Generator: high-res response
Generator->>FileSystem: write final image
FileSystem-->>Generator: file path
Generator-->>CLI: return Path
CLI-->>User: finalized image path
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes
Suggested labels
Poem
Pre-merge checks and finishing touches✅ Passed checks (3 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
✅ FIPS Compatibility Check
Status: ✅ PASSED What is FIPS?FIPS 140-2/140-3 is a US government standard for cryptographic modules. Common issues:
|
coderabbitai
Bot
added
dependencies
documentation
✅ Mutation Testing Results
What is Mutation Testing?Mutation testing introduces small changes (mutations) to your code and checks if your tests detect them. A high mutation score indicates your tests are effective at catching bugs.
|
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 11
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: ASSERTIVE
Plan: Pro
⛔ Files ignored due to path filters (1)
uv.lockis excluded by!**/*.lock,!**/*.lock
📒 Files selected for processing (12)
packages/gemini-image/README.md(1 hunks)packages/gemini-image/pyproject.toml(1 hunks)packages/gemini-image/src/gemini_image/__init__.py(1 hunks)packages/gemini-image/src/gemini_image/cli.py(1 hunks)packages/gemini-image/src/gemini_image/generator.py(1 hunks)packages/gemini-image/src/gemini_image/models.py(1 hunks)packages/gemini-image/src/gemini_image/utils.py(1 hunks)packages/gemini-image/tests/conftest.py(1 hunks)packages/gemini-image/tests/test_generator.py(1 hunks)packages/gemini-image/tests/test_models.py(1 hunks)packages/gemini-image/tests/test_utils.py(1 hunks)pyproject.toml(4 hunks)
🧰 Additional context used
📓 Path-based instructions (3)
**/*.py
📄 CodeRabbit inference engine (CLAUDE.md)
**/*.py: Use Ruff formatting with 88 character line length for Python code
Use Ruff linting with PyStrict-aligned rules including BLE, EM, SLF, INP, ISC, PGH, RSE, TID, YTT, FA, T10, and G rules
Tag assumptions with #CRITICAL, #ASSUME, or #EDGE comments including category and reason for verification
Files:
packages/gemini-image/src/gemini_image/utils.pypackages/gemini-image/src/gemini_image/cli.pypackages/gemini-image/tests/test_utils.pypackages/gemini-image/tests/test_models.pypackages/gemini-image/tests/conftest.pypackages/gemini-image/src/gemini_image/__init__.pypackages/gemini-image/tests/test_generator.pypackages/gemini-image/src/gemini_image/models.pypackages/gemini-image/src/gemini_image/generator.py
**/*.md
📄 CodeRabbit inference engine (CLAUDE.md)
Use 120 character line length for Markdown documentation files
Files:
packages/gemini-image/README.md
pyproject.toml
⚙️ CodeRabbit configuration file
pyproject.toml: Review dependency changes for:
- Version constraint appropriateness
- Security implications of new dependencies
- License compatibility
Files:
pyproject.toml
🧠 Learnings (2)
📚 Learning: 2025-12-14T22:54:23.007Z
Learnt from: CR
Repo: ByronWilliamsCPA/python-libs PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-14T22:54:23.007Z
Learning: Applies to tests/**/*.py : Use pytest fixtures defined in tests/conftest.py for test setup and teardown
Applied to files:
packages/gemini-image/tests/conftest.py
📚 Learning: 2025-12-14T22:54:23.007Z
Learnt from: CR
Repo: ByronWilliamsCPA/python-libs PR: 0
File: CLAUDE.md:0-0
Timestamp: 2025-12-14T22:54:23.007Z
Learning: Applies to src/**/*.py : Use BasedPyright for type checking in strict mode with strict inference enabled
Applied to files:
pyproject.toml
🧬 Code graph analysis (4)
packages/gemini-image/tests/test_utils.py (2)
packages/gemini-image/src/gemini_image/utils.py (4)
decode_base64_image(81-91)get_api_key(13-44)get_file_extension(94-110)load_image_as_base64(47-78)packages/gemini-image/tests/conftest.py (2)
sample_image_path(26-30)sample_image_bytes(16-22)
packages/gemini-image/src/gemini_image/__init__.py (2)
packages/gemini-image/src/gemini_image/generator.py (2)
generate_image(55-304)generate_story_sequence(307-428)packages/gemini-image/src/gemini_image/models.py (1)
ModelConfig(13-19)
packages/gemini-image/tests/test_generator.py (2)
packages/gemini-image/src/gemini_image/generator.py (3)
finalize_draft(431-501)generate_story_sequence(307-428)generate_image(55-304)packages/gemini-image/tests/conftest.py (2)
mock_genai_response(34-51)sample_image_path(26-30)
packages/gemini-image/src/gemini_image/generator.py (1)
packages/gemini-image/src/gemini_image/utils.py (3)
get_api_key(13-44)get_file_extension(94-110)load_image_as_base64(47-78)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Mutation Testing / Mutation Testing
🔇 Additional comments (16)
pyproject.toml (2)
271-271: LGTM! Proper monorepo integration.The
gemini_imagepackage is correctly added to:
- Ruff's
known-first-partyfor import sorting- BasedPyright's include paths for type checking
- Pytest's
testpathsandpythonpathfor test discoveryThis ensures consistent tooling across the monorepo.
Also applies to: 465-465, 528-529
700-703: [rewritten comment]
[classification tag]packages/gemini-image/pyproject.toml (3)
1-66: Well-structured package configuration.The pyproject.toml follows best practices:
- Comprehensive metadata with classifiers and keywords
- Clear Python version constraints
- Proper semantic release configuration with package-specific tag format
- Hatchling build configuration
26-26: No action needed —google-genai>=1.0.0is valid.The
google-genaipackage exists on PyPI at version 1.55.0, which satisfies the specified requirement. The package name is correct and unambiguous.
25-27: The CLI implementation usesargparse(standard library), notclick. No missing dependency here.Likely an incorrect or invalid review comment.
packages/gemini-image/tests/test_utils.py (1)
1-111: Good test coverage for utility functions.The tests comprehensively cover:
- API key retrieval from environment and
.envfiles- Image loading with MIME type detection
- Base64 encoding/decoding round-trip
- File extension mapping with fallback behavior
Test structure follows pytest best practices with clear class organization and descriptive method names.
packages/gemini-image/README.md (1)
99-104: Default model documentation is correct.The
DEFAULT_MODELingemini_image/models.pyis set to"pro", which matches the table indicatingproas the default model.packages/gemini-image/src/gemini_image/cli.py (1)
1-230: Documentation inconsistency: CLI uses argparse, not Click.The PR summary states "Click-based command-line interface," but the implementation uses
argparse(imported on line 5). This documentation inconsistency should be corrected in the PR description.packages/gemini-image/tests/conftest.py (1)
1-59: LGTM!The pytest fixtures are well-structured and provide appropriate mocks for testing the Gemini image generation functionality. The fixtures correctly simulate the API response structure and follow pytest best practices.
packages/gemini-image/tests/test_models.py (1)
1-38: LGTM!The test coverage for model configurations is appropriate and validates all the key constants used throughout the package.
packages/gemini-image/src/gemini_image/utils.py (2)
47-78: LGTM!The
load_image_as_base64function correctly handles image loading, MIME type detection, and base64 encoding with proper error handling for missing files.
81-110: LGTM!The
decode_base64_imageandget_file_extensionutility functions are simple, correct, and follow best practices.packages/gemini-image/src/gemini_image/generator.py (3)
1-52: LGTM!The lazy loading pattern for the google.genai dependency is well-implemented and provides clear error messages. The complexity warning suppressions are appropriately documented in the module docstring.
307-428: LGTM with note on type handling.The
generate_story_sequencefunction correctly implements multi-part story generation with visual continuity. The function properly validates inputs, builds appropriate prompts for each part, and uses the previous image as a reference.Note: The function assumes
output_prefixis always a Path object at line 386 (.stemaccess), which is safe because line 349 ensures it's a Path when None is provided. The type hint correctly reflects this.
431-501: LGTM!The
finalize_draftfunction correctly implements draft finalization by delegating togenerate_imagewith appropriate parameters. The function properly validates the draft exists and provides sensible defaults for aspect ratio and image size.packages/gemini-image/src/gemini_image/models.py (1)
22-37: The model IDs"gemini-2.5-flash-image"and"gemini-3-pro-image-preview"are valid and confirmed in official Google Gemini API documentation for image generation. No changes needed.
| A comprehensive image generation library built on Google's Gemini models (Nano Banana / Nano Banana Pro). | ||
|
|
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Clarify or remove internal codenames.
"Nano Banana / Nano Banana Pro" appears to be internal codenames that may confuse end users. Consider either removing these or adding context explaining what they refer to.
🤖 Prompt for AI Agents
In packages/gemini-image/README.md around lines 3-4, the README references
internal codenames "Nano Banana / Nano Banana Pro" which may confuse users;
either remove these codenames or add a short parenthetical explaining they are
internal model nicknames (or map them to their official public model names),
update the sentence accordingly to use the public/official model names or
include a brief clarifying phrase, and ensure the README remains clear and
user-facing.
| # Multi-part story sequence | ||
| from gemini_image import generate_story_sequence | ||
|
|
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Remove duplicate import statement.
generate_story_sequence is already imported on line 34. This duplicate import in the code example is redundant and may confuse readers.
# Multi-part story sequence
-from gemini_image import generate_story_sequence
-
images = generate_story_sequence(🤖 Prompt for AI Agents
In packages/gemini-image/README.md around lines 61 to 63, the example duplicates
the import of generate_story_sequence (it's already imported on line 34); remove
the duplicate import line so the example uses the previously declared import, or
consolidate both examples to reference the single import at the top to avoid
redundancy.
| from gemini_image.generator import generate_image, generate_story_sequence | ||
| from gemini_image.models import ( | ||
| ASPECT_RATIOS, | ||
| DEFAULT_MODEL, | ||
| IMAGE_SIZES, | ||
| MODELS, | ||
| AspectRatio, | ||
| ImageSize, | ||
| ModelConfig, | ||
| ModelKey, | ||
| ) | ||
|
|
||
| __all__ = [ | ||
| "ASPECT_RATIOS", | ||
| "DEFAULT_MODEL", | ||
| "IMAGE_SIZES", | ||
| "MODELS", | ||
| "AspectRatio", | ||
| "ImageSize", | ||
| "ModelConfig", | ||
| "ModelKey", | ||
| "generate_image", | ||
| "generate_story_sequence", | ||
| ] |
There was a problem hiding this comment.
Missing finalize_draft export.
The README.md documents finalize_draft() in the API Reference section (lines 172-185), but this function is not exported from __init__.py. Users following the documentation won't be able to import it from the package namespace.
Apply this diff to export finalize_draft:
-from gemini_image.generator import generate_image, generate_story_sequence
+from gemini_image.generator import finalize_draft, generate_image, generate_story_sequenceAnd update __all__:
__all__ = [
"ASPECT_RATIOS",
"DEFAULT_MODEL",
"IMAGE_SIZES",
"MODELS",
"AspectRatio",
"ImageSize",
"ModelConfig",
"ModelKey",
+ "finalize_draft",
"generate_image",
"generate_story_sequence",
]🤖 Prompt for AI Agents
In packages/gemini-image/src/gemini_image/__init__.py around lines 23 to 46, the
function finalize_draft is documented in the README but not exported from the
package; import finalize_draft from its module (likely from
gemini_image.generator or the correct module where it's defined) at the top of
the file and add "finalize_draft" to the __all__ list so it is available from
the package namespace for users following the docs.
| if args.story_parts < 2: | ||
| print("Error: Story must have at least 2 parts") # noqa: T201 | ||
| sys.exit(1) |
There was a problem hiding this comment.
Story parts validation should match generator's requirement.
The CLI validates args.story_parts < 2 (line 183), requiring at least 2 parts, but generate_story_sequence in generator.py validates num_parts < 1 (line 340), requiring at least 1 part. This inconsistency could confuse users who might legitimately want a single-part story.
Apply this diff to align the validation:
# Story sequence mode
if args.story_parts:
- if args.story_parts < 2:
- print("Error: Story must have at least 2 parts") # noqa: T201
+ if args.story_parts < 1:
+ print("Error: Story must have at least 1 part") # noqa: T201
sys.exit(1)📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| if args.story_parts < 2: | |
| print("Error: Story must have at least 2 parts") # noqa: T201 | |
| sys.exit(1) | |
| if args.story_parts < 1: | |
| print("Error: Story must have at least 1 part") # noqa: T201 | |
| sys.exit(1) |
🤖 Prompt for AI Agents
In packages/gemini-image/src/gemini_image/cli.py around lines 183 to 185, the
CLI enforces a minimum of 2 story parts while generate_story_sequence expects at
least 1; update the CLI validation to match the generator by changing the check
to require args.story_parts < 1 and adjust the error message to "Error: Story
must have at least 1 part" (and exit with sys.exit(1)) so the two validations
are consistent.
| img_data, mime_type = load_image_as_base64(img_path) | ||
| contents.append( | ||
| types.Part.from_bytes( | ||
| data=base64.standard_b64decode(img_data), | ||
| mime_type=mime_type, | ||
| ) | ||
| ) |
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Inefficient encoding/decoding roundtrip.
Lines 122-125 encode the image to base64 via load_image_as_base64, then immediately decode it back to bytes with base64.standard_b64decode(img_data). This is inefficient.
Consider one of these approaches:
- Option 1: Load the binary data directly without the encoding/decoding roundtrip:
with open(img_path, "rb") as f:
img_bytes = f.read()
# Determine mime_type from suffix
suffix = img_path.suffix.lower()
mime_types = {".png": "image/png", ".jpg": "image/jpeg", ...}
mime_type = mime_types.get(suffix, "image/png")
contents.append(
types.Part.from_bytes(
data=img_bytes,
mime_type=mime_type,
)
)- Option 2: Add a helper function in utils.py to load images as bytes directly, and use that instead of
load_image_as_base64here.
🤖 Prompt for AI Agents
In packages/gemini-image/src/gemini_image/generator.py around lines 122 to 128,
the code currently calls load_image_as_base64 and then immediately decodes it
back to bytes, causing an unnecessary base64 roundtrip; replace that with direct
binary reads (or call a new helper that returns bytes) and determine mime_type
from the file suffix (use a small suffix->mime map with sensible default) before
calling types.Part.from_bytes with the raw bytes and mime_type.
| if not api_key and env_file and env_file.exists(): | ||
| with open(env_file) as f: | ||
| for line in f: | ||
| line = line.strip() | ||
| if line.startswith("GEMINI_API_KEY="): | ||
| api_key = line.split("=", 1)[1].strip().strip('"').strip("'") | ||
| break |
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Specify encoding when opening text files.
Line 30 opens the .env file without specifying an encoding parameter. For text files, it's best practice to explicitly specify the encoding (typically 'utf-8') to ensure consistent behavior across platforms.
Apply this diff:
if not api_key and env_file and env_file.exists():
- with open(env_file) as f:
+ with open(env_file, encoding="utf-8") as f:
for line in f:
line = line.strip()
if line.startswith("GEMINI_API_KEY="):
api_key = line.split("=", 1)[1].strip().strip('"').strip("'")
break📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| if not api_key and env_file and env_file.exists(): | |
| with open(env_file) as f: | |
| for line in f: | |
| line = line.strip() | |
| if line.startswith("GEMINI_API_KEY="): | |
| api_key = line.split("=", 1)[1].strip().strip('"').strip("'") | |
| break | |
| if not api_key and env_file and env_file.exists(): | |
| with open(env_file, encoding="utf-8") as f: | |
| for line in f: | |
| line = line.strip() | |
| if line.startswith("GEMINI_API_KEY="): | |
| api_key = line.split("=", 1)[1].strip().strip('"').strip("'") | |
| break |
🤖 Prompt for AI Agents
In packages/gemini-image/src/gemini_image/utils.py around lines 29 to 35, the
.env file is opened without an explicit encoding; update the open call to
specify encoding='utf-8' (e.g., with open(env_file, 'r', encoding='utf-8') as f)
so the file is read with a consistent text encoding across platforms while
preserving the existing logic.
| patch.object(generator, "_types", mock_types), | ||
| patch.dict(os.environ, {}, clear=True), | ||
| ): | ||
| os.environ.pop("GEMINI_API_KEY", None) |
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Redundant operation after clear=True.
After patch.dict(os.environ, {}, clear=True) on line 44, calling os.environ.pop("GEMINI_API_KEY", None) on line 46 is redundant since the environment has already been cleared.
Apply this diff to remove the redundant line:
with (
patch.object(generator, "_genai", mock_genai),
patch.object(generator, "_types", mock_types),
patch.dict(os.environ, {}, clear=True),
):
- os.environ.pop("GEMINI_API_KEY", None)
with pytest.raises(ValueError, match="GEMINI_API_KEY"):
generator.generate_image("test prompt")📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| os.environ.pop("GEMINI_API_KEY", None) | |
| with ( | |
| patch.object(generator, "_genai", mock_genai), | |
| patch.object(generator, "_types", mock_types), | |
| patch.dict(os.environ, {}, clear=True), | |
| ): | |
| with pytest.raises(ValueError, match="GEMINI_API_KEY"): | |
| generator.generate_image("test prompt") |
🤖 Prompt for AI Agents
In packages/gemini-image/tests/test_generator.py at line 46, the call
os.environ.pop("GEMINI_API_KEY", None) is redundant because
patch.dict(os.environ, {}, clear=True) on line 44 already clears the
environment; remove the os.environ.pop(...) line so the redundant operation is
eliminated and the test relies on the cleared environment.
| results = generator.generate_story_sequence( | ||
| base_prompt="A test story", | ||
| num_parts=3, | ||
| output_dir=tmp_path, | ||
| output_prefix=tmp_path / "story", | ||
| ) |
There was a problem hiding this comment.
Type mismatch: output_prefix expects a prefix string, not a full Path.
Line 138 passes tmp_path / "story" as output_prefix, but according to the generator implementation (line 386 in generator.py), output_prefix is used as Path(f"{output_prefix.stem}_part{part_num}.png"). This works but is semantically incorrect—output_prefix should be a simple string prefix (e.g., "story"), not a full path that includes the directory.
Apply this diff to fix the semantic issue:
results = generator.generate_story_sequence(
base_prompt="A test story",
num_parts=3,
output_dir=tmp_path,
- output_prefix=tmp_path / "story",
+ output_prefix=Path("story"),
)📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| results = generator.generate_story_sequence( | |
| base_prompt="A test story", | |
| num_parts=3, | |
| output_dir=tmp_path, | |
| output_prefix=tmp_path / "story", | |
| ) | |
| results = generator.generate_story_sequence( | |
| base_prompt="A test story", | |
| num_parts=3, | |
| output_dir=tmp_path, | |
| output_prefix=Path("story"), | |
| ) |
🤖 Prompt for AI Agents
In packages/gemini-image/tests/test_generator.py around lines 134 to 139, the
test passes tmp_path / "story" (a Path) as output_prefix but the generator
expects a simple prefix string; change the call to pass output_prefix="story"
and keep output_dir=tmp_path (i.e., replace tmp_path / "story" with "story") so
the generator composes filenames correctly; update any related assertions in the
test that assumed a path-based prefix to reflect the string-based prefix usage.
| with patch.dict(os.environ, {}, clear=True): | ||
| # Remove GEMINI_API_KEY if it exists | ||
| os.environ.pop("GEMINI_API_KEY", None) | ||
| assert get_api_key(env_file=env_file) == "file-key-456" |
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Redundant os.environ.pop after clear=True.
When patch.dict(os.environ, {}, clear=True) is used, the environment is already cleared. The subsequent os.environ.pop("GEMINI_API_KEY", None) is redundant.
def test_get_api_key_from_env_file(self, tmp_path: Path) -> None:
"""Test getting API key from .env file."""
env_file = tmp_path / ".env"
env_file.write_text('GEMINI_API_KEY="file-key-456"')
with patch.dict(os.environ, {}, clear=True):
- # Remove GEMINI_API_KEY if it exists
- os.environ.pop("GEMINI_API_KEY", None)
assert get_api_key(env_file=env_file) == "file-key-456"🤖 Prompt for AI Agents
In packages/gemini-image/tests/test_utils.py around lines 36 to 39, the test
uses patch.dict(os.environ, {}, clear=True) which already clears the
environment, so the subsequent os.environ.pop("GEMINI_API_KEY", None) is
redundant; remove that pop call (or the entire redundant line) so the test
relies solely on the cleared env and still asserts
get_api_key(env_file=env_file) == "file-key-456".
| with patch.dict(os.environ, {}, clear=True): | ||
| os.environ.pop("GEMINI_API_KEY", None) | ||
| with pytest.raises(ValueError, match="GEMINI_API_KEY"): | ||
| get_api_key() |
There was a problem hiding this comment.
🧹 Nitpick | 🔵 Trivial
Same redundancy here.
The os.environ.pop is unnecessary after clear=True.
def test_get_api_key_missing_raises(self) -> None:
"""Test that missing API key raises ValueError."""
with patch.dict(os.environ, {}, clear=True):
- os.environ.pop("GEMINI_API_KEY", None)
with pytest.raises(ValueError, match="GEMINI_API_KEY"):
get_api_key()🤖 Prompt for AI Agents
In packages/gemini-image/tests/test_utils.py around lines 43 to 46, the test
uses patch.dict(os.environ, {}, clear=True) and then redundantly calls
os.environ.pop("GEMINI_API_KEY", None); remove the os.environ.pop line since
clear=True already empties the environment, leaving only the patch.context and
the pytest.raises assertion to verify get_api_key() raises a ValueError
mentioning "GEMINI_API_KEY".
|
1 participant
Summary
Adds a new
gemini-imagepackage to the monorepo for generating images using Google's Gemini API.Features
Package Structure
gemini_image.generator- Core image generation functionsgemini_image.models- Model configurations and constantsgemini_image.utils- Utility functions (API key handling, image encoding)gemini_image.cli- Click-based command-line interfaceTesting
Test plan
🤖 Generated with Claude Code
Summary by CodeRabbit
New Features
Documentation
✏️ Tip: You can customize this high-level summary in your review settings.