Add Prompt Templates

README.md

@@ -50,6 +50,12 @@ Welcome to **FriendlyMeals**, an app that demonstrates how to integrate Firebase
      - Go to the [**App Check**](https://console.firebase.google.com/project/_/appcheck/?useAutoProject=true) section of the console, enable this service, and register your Android app using the Play Integrity provider (you'll need to provide the SHA-256 fingerprint of your app's signing certificate).
      - Go to the [**Remote Config**](https://console.firebase.google.com/project/_/config/?useAutoProject=true) section of the console and enable this service.
      - Go to the [**Firebase Storage**](https://console.firebase.google.com/project/_/storage/?useAutoProject=true) section of the console and enable this service (you'll need to set up a billing account since this service is only available on the Blaze plan).
+
+1. **Create Firebase AI Logic Prompt Templates**
+     - If you're not familiar with Prompt templates, take a look at the [documentation](https://firebase.google.com/docs/ai-logic/server-prompt-templates/get-started) to learn more about it.
+     - Go to the [AI services > AI Logic > Prompt templates tab](https://console.firebase.google.com/project/_/ailogic/templates/?useAutoProject=true).
+     - Create one prompt template for each one of the four templates in the [prompts folder](https://github.com/FirebaseExtended/FriendlyMeals-Android/tree/main/prompts). Please note that the model names below might not be the most up to date ones. [Refer to the documentation](https://firebase.google.com/docs/ai-logic/models) to check the latest available models.
+     - Don't forget to test each one of your templates in the Firebase console before you use them in your apps. [Read the docs](https://firebase.google.com/docs/ai-logic/server-prompt-templates/get-started?api=dev#test-template) to learn how to test your templates properly.
 
 1. **Run the app**
    - Build and run in Android Studio on an Android emulator or physical device.

prompts/find-stores-template-v1-0-0.prompt

@@ -0,0 +1,18 @@
+---
+model: "gemini-3.5-flash-lite"
+tools:
+  - googleMaps
+input:
+  schema:
+    ingredients: "string, the ingredients in the shopping list"
+    dayOfWeek: "string, current day of the week"
+    currentTime: "string, current time of the day"
+---
+
+What are the nearest grocery stores or markets near me that stock these ingredients: {{ingredients}}? 
+For each place, tell me their business hours, if it's open right now on {{dayOfWeek}} at {{currentTime}}, and if it's closing in less than 30 minutes. 
+Tell me what the parking situation is like at each place: is there a dedicated lot or should I look for street parking? 
+Tell the Map URL so I can open it in Google Maps. 
+Format your response strictly as a JSON object with a "stores" array containing store objects. 
+Each store object must have these EXACT keys: "name": string, "address": string, "distance": string, "openNow": boolean, "closingSoon": boolean, "hasParking": boolean, "parkingDetails": string, "mapUrl": string. 
+Do NOT include markdown code block formatting (like ```json). Output only the raw JSON string.

prompts/generate-recipe-photo-gemini-template-v1-0-0.prompt

@@ -0,0 +1,9 @@
+---
+model: "gemini-3.1-flash-image"
+input:
+  schema:
+    recipeTitle: "string"
+---
+
+A professional food photography shot of this recipe: {{recipeTitle}}. 
+Style: High-end food photography, restaurant-quality plating, soft natural lighting, on a clean background, showing the complete plated dish.

prompts/generate-recipe-template-v1-0-0.prompt

@@ -0,0 +1,30 @@
+---
+model: "gemini-3.5-flash-lite"
+input:
+  schema:
+    ingredients: "string, the ingredients list"
+    notes?: "string, cuisine and dietary notes"
+output:
+  format: json
+  schema:
+    title: string
+    instructions: string
+    ingredients(array, ingredients for recipe): string
+    prepTime: string
+    cookTime: string
+    servings: string
+    tags(array, relevant tags for recipe): string
+---
+
+{{role "user"}}
+Create a detailed recipe based on these ingredients: {{ingredients}}.
+
+Format requirements:
+ - 'instructions': Provide the cooking steps as a clear list of instructions separated by newlines. Use bold formatting on the step numbers. Use Markdown.
+ - 'ingredients': List all necessary items, including quantities.
+ - 'prepTime', 'cookTime', 'servings': Short strings (e.g., "15 mins").
+ - 'tags': Generate a list of 3-5 relevant category tags (e.g., "Healthy", "Vegan", "Gluten-Free", "Dessert", "Quick").
+
+{{#if notes}}
+IMPORTANT CUISINE AND DIETARY NOTES: {{notes}}
+{{/if}}

prompts/scan-meal-template-v1-0-0.prompt

@@ -0,0 +1,16 @@
+---
+model: "gemini-3.5-flash-lite"
+output:
+  format: json
+  schema:
+    protein: string
+    fat: string
+    carbs: string
+    sugar: string
+    ingredients(array, ingredients in the meal): string
+---
+
+Analyze this image of a meal: {{media type="mimeType" data="imageData"}}, and estimate the nutritional content.
+Return the result in JSON format matching the schema:
+- protein, fat, carbs, sugar (strings with units, e.g., '20g')
+- ingredients (list of strings)