Calorie Journal Factory

What problem it solves

Daily meal logging often fails at three practical points: counting calories is tedious, writing full journal entries feels heavy, and nutrition numbers can create unnecessary pressure. Calorie Journal Factory targets this lightweight logging use case. It turns natural-language meal notes into copyable text journals or PNG cards, with mood tagging, animal matching, and data checks, so the output feels more like a daily summary than a performance report.

How it works

• Input: describe meals, snacks, exercise, and basic context; the more specific the food names, quantities, and cooking methods are, the more stable the result.
• Recognition and estimation: search_engine_v2.py runs a multi-level food lookup, preferring extension data, API results, and built-in data, with estimation as the fallback.
• Data checks: gen_card.py validates seven dimensions before card generation to reduce obvious calorie, protein, and carbohydrate errors.
• Dual output: text_handbook_generator.py emits Markdown or plain text, while gen_card.py renders image cards, typically covering today’s meals, mood review, animal match, and summary pages.
• Themes and stickers: theme_pack_loader.py loads JSON theme packs across six modules: naming, character, mood, tone, visual style, and stickers; users can use default, space_explorer, or a custom theme.
• Matching logic: food_mood_analyzer.py and animal_blind_box.py match results using nutrition ratios, snacks, calorie levels, and related conditions rather than random decoration.

Boundaries and cautions

It fits everyday meals, common Chinese home dishes, mood-oriented logging, and lightweight sharing. It is not suitable for medical diets for diabetes, kidney disease, or other clinical conditions, nor does it replace professional nutrition calculation or medical diagnosis. For very large one-day logs, split entries by day. Vague inputs may be marked as estimates. If structured nutrition data already exists and only needs formatting, a document tool is usually more direct. Outputs are for personal reference only.