Added cost calculation for Image models

[ViaSocket-Git]

added the cost calculation for the image models
there were different type of calculations for the openai , gemini, imagen

[ViaSocket-Git]

/windsurf-review

[windsurf-bot]

Other comments (6)

• src/services/commonServices/Google/gemini_image_utils.py (101-101) The `response.model_dump()` call might fail if the response object doesn't have this method. Consider adding a try-except block or checking if the method exists before calling it:

              "usage_metadata": getattr(response, 'model_dump', lambda: {})().get('usage_metadata', {}) | {"total_images_generated": (len(gcp_urls))}
  

• src/services/commonServices/common.py (609-615) The image cost calculation code doesn't include error handling. If `token_calculator.calculate_image_usage()` or `calculate_image_cost()` fails (e.g., for unsupported models), it will cause an unhandled exception. Consider adding try/except blocks to gracefully handle potential calculation errors.
• src/services/commonServices/common.py (615-615) The code assigns `total_cost` to the response without validating it. Consider adding a fallback value or validation to ensure `parsed_data['tokens'].get('total_cost')` returns a valid number to avoid potential issues with downstream code that expects a numeric value.
• src/services/utils/ai_middleware_format.py (181-182) The 'revised_prompt' field was moved from individual image items to the parent data object. This creates an inconsistency with how it's structured in the else branch and may break clients expecting this field in both places.
• src/services/commonServices/Google/gemini_image_utils.py (101-101) The parentheses around `len(gcp_urls)` are unnecessary and can be removed for cleaner code:

              "usage_metadata": response.model_dump().get('usage_metadata', {}) | {"total_images_generated": len(gcp_urls)}
  

• src/services/utils/token_calculation.py (175-175) The docstring for `calculate_total_cost` states it "Automatically detects image models and uses appropriate calculation", but there's no actual detection logic in the method. Either implement this detection logic or update the docstring to reflect the actual behavior.

💡 To request another review, post a new comment with "/windsurf-review".

[windsurf-bot]

The 'url' variable was removed on line 175, but it's still being referenced in the else branch on lines 198-199. This will cause a NameError when processing a single image response from Gemini.

[ViaSocket-Git]

there is no issue like this. its just a comment

[windsurf-bot]

Similar to the input tokens, the output token calculation for Gemini should accumulate values rather than overwrite them:

Suggested change

	if modality == 'TEXT':
	usage['text_output_tokens'] = token_count
	elif modality == 'IMAGE':
	usage['image_output_tokens'] = token_count
	if modality == 'TEXT':
	usage['text_output_tokens'] = usage.get('text_output_tokens', 0) + token_count
	elif modality == 'IMAGE':
	usage['image_output_tokens'] = usage.get('image_output_tokens', 0) + token_count

[ViaSocket-Git]

no, there will be no multiple entries with the same modality

[Husainbw786]

explained review chnages do it ASAP