Weather Outfit Advisor

Overview

Provides personalized outfit recommendations based on real-time weather data for your destination and your personal clothing preferences.

Workflow

Step 1: Check Information Completeness

Before providing recommendations, must confirm the following three key pieces of information:

1. 1. Destination (city name)
2. Date (specific date or date range)
3. Clothing Preferences (style preference, temperature sensitivity, special needs, etc.)

Information Extraction Strategy:

• - Extract explicit destination, date, and preferences from user input
• Recognize relative date terms ("tomorrow", "day after tomorrow", "next Monday", etc.) and convert to specific dates
• If any item is missing, must ask the user first

Relative Date Conversion Rules:

• - "tomorrow" → current date + 1 day
• "day after tomorrow" → current date + 2 days
• "three days later" → current date + 3 days
• "next Monday" → next Monday's date
• "this weekend" → this Saturday or Sunday (needs confirmation)

Example:
CODEBLOCK0

Step 2: Collect Necessary Information

Adopt different clarification strategies based on the number of missing items:

Strategy A: Missing Only 1 Item (Targeted Question)

Ask only about the missing item:

If only missing clothing preferences:
CODEBLOCK1

If only missing date:
CODEBLOCK2

If only missing destination:
CODEBLOCK3

Strategy B: Missing 2-3 Items (Ask All at Once)

When multiple items are missing, list all questions at once:
CODEBLOCK4

Relative Date Handling

When users use relative dates:

Recognize Keywords:

• - tomorrow, day after tomorrow, three days later
• next Monday, next Tuesday... next Sunday
• this weekend, next weekend
• next week, next month

Conversion Logic:

1. 1. Get current system date
2. Calculate specific date based on relative term
3. Confirm the converted date in the response

Example:
CODEBLOCK5

Step 3: Query Weather Data

Use the provided Python script (recommended):

CODEBLOCK6

Script Features:

• - ✅ Automatic relative date parsing (tomorrow, day after tomorrow, next Monday, etc.)
• ✅ City name normalization (supports Chinese/English and common aliases)
• ✅ Calls wttr.in API to get weather data
• ✅ Returns structured JSON data
• ✅ Includes current weather and forecast information

Output Example (JSON format):
CODEBLOCK7

Extract Key Information from Output:
CODEBLOCK8

Alternative: Direct API Call (if script is unavailable):

Current Weather API:
CODEBLOCK9

Example:
CODEBLOCK10

Parse JSON Response for Key Information:
CODEBLOCK11

Alternative: OpenWeatherMap API (requires free API key):

If wttr.in is unavailable, you can use OpenWeatherMap:

API Endpoint:
CODEBLOCK12

Forecast API:
CODEBLOCK13

Note:

• - {API_KEY} needs to be requested from https://openweathermap.org/api (free version is sufficient)
• INLINECODE1 uses Celsius
• INLINECODE2 returns English description

Step 4: Analyze Weather Data

Extract key information from the API response:

CODEBLOCK14

Step 5: Generate Outfit Recommendations

Generate recommendations based on weather data and user preferences:

Temperature Range Reference

Temperature Range	Outfit Recommendation
< 5°C	Heavy down jacket, thermal underwear, sweater, scarf, gloves, hat
5-10°C

Thick coat, padded jacket, knitwear, thin sweater | | 10-15°C | Trench coat, jacket, hoodie, long-sleeve T-shirt | | 15-20°C | Light jacket, denim jacket, long-sleeve shirt, T-shirt + cardigan | | 20-25°C | Long-sleeve T-shirt, thin shirt, single layer clothing | | 25-30°C | Short-sleeve T-shirt, shirt, skirt, shorts | | > 30°C | Breathable short-sleeve, tank top, sun protection clothing, sun hat |

Consider Other Factors

Rainy Days:

• - Carry umbrella or raincoat
• Choose waterproof shoes
• Avoid light-colored clothing

Windy Days:

• - Avoid skirts or choose anti-exposure styles
• Wear windproof jacket
• Secure hat properly

High Humidity:

• - Choose breathable quick-dry fabrics
• Avoid heavy clothing

Strong UV:

• - Apply sunscreen
• Wear sunglasses
• Wear sun protection clothing or long sleeves

Add Image Reference (Optional but Recommended)

After providing outfit recommendations, call the image search script to provide visual references:

CODEBLOCK15

Examples:
CODEBLOCK16

Step 6: Output Format

Output recommendations in the following format:

CODEBLOCK17

Special Cases Handling

API Call Failure

If API call fails or returns error:

CODEBLOCK18

Extreme Weather Warning

If extreme weather is detected (typhoon, blizzard, heavy rain warning, etc.):

CODEBLOCK19

Multi-Day Trip

If user provides a date range:

CODEBLOCK20

Personalization Adjustments

Adjust recommendations based on user's clothing habits:

Cold-sensitive users: Recommend adding extra warm layers to base outfit
Heat-sensitive users: Recommend more breathable fabrics, reduce layers
Business needs: Maintain professional appearance while ensuring comfort
Outdoor sports: Prioritize functional clothing (quick-dry, waterproof, windproof)
Pregnant/elderly/children: Pay extra attention to warmth and comfort

Example Dialogues

Example 1: Complete Information

User: I'm going to Hangzhou tomorrow, I'm usually sensitive to cold, prefer casual style.

Assistant: Step 1: Query weather
CODEBLOCK21

Step 2: Parse data
Extract from returned JSON:

• - Temperature range: 11-21°C
• Weather condition: Patchy rain nearby
• Humidity: 56%
• Rain chance: 70%

Step 3: Generate recommendations

According to weather forecast, Hangzhou tomorrow:

• - Temperature: 11-21°C
• Weather: Patchy rain nearby
• Humidity: 56%

Outfit recommendations:

• - Top: Long-sleeve T-shirt + knit cardigan + thin trench coat (considering you're sensitive to cold, recommend bringing thermal underwear as backup)
• Bottom: Jeans or casual pants
• Shoes: Sneakers or casual shoes
• Accessories: Can bring a thin scarf, umbrella

Example 2: Incomplete Information

User: I'm going out to play in a few days, how should I dress?

Assistant: To provide you with the most suitable outfit recommendations, I need some information:

1. 1. What is your travel destination?
2. When exactly are you traveling?
3. What are your usual clothing habits? Any special preferences (such as sensitive to cold/heat, prefer casual/business style, etc.)?

Example 3: Real Call Case

User: I'm going to Hangzhou tomorrow, I'm quite heat-sensitive

Assistant: Execution Flow:

1. 1. Information Extraction:

- Destination: Hangzhou ✓
- Date: tomorrow (relative date) → convert to specific date
- Clothing preferences: heat-sensitive ✓

1. 2. Date Conversion:

- Get current system date: 2026-04-01 - Calculate: tomorrow = 2026-04-02 - Confirm: "Sure, I'll check the weather for Hangzhou tomorrow (April 2nd)"

1. 3. Use wttr.in API to query Hangzhou weather: INLINECODE3

1. 4. Parse JSON to get data:

- Current temperature: 19°C - Max temp: 21°C, Min temp: 11°C - Weather: Patchy rain nearby - Humidity: 56%

1. 5. Adjust recommendations based on heat-sensitive preference

Then output complete outfit recommendations...

Example 4: Boundary Case Testing

Scenario A: Seriously Insufficient Information

User: How should I dress?

Assistant: To provide you with the most suitable outfit recommendations, I need some information:

1. 1. What is your travel destination?
2. When exactly are you traveling? (You can say "tomorrow", "April 5th", etc.)
3. What are your usual clothing habits? (For example, sensitive to cold/heat, preferred style, etc.)

Scenario B: Only Partial Information

User: I'm going to Shanghai the day after tomorrow

Assistant: Sure, I'll help you check the weather for Shanghai the day after tomorrow.

To make the recommendation more suitable for you, please tell me about your clothing preferences:

• - Are you usually sensitive to cold or heat?
• What style do you prefer? (casual/business/sports, etc.)
• Any special needs? (such as formal occasions, outdoor activities, etc.)

(Wait for user response, then query weather and provide complete recommendations)

Scenario C: Multi-Day Trip

User: I'm going to Beijing on business from next Monday to Friday, I'm sensitive to cold

Assistant: Execution Flow:

1. 1. Recognize this is a date range: next Monday to Friday (total 5 days)
2. Query weather forecast for each of the 5 days separately
3. Provide daily outfit recommendations
4. Give overall suggestions for clothing packing list (considering laundry and matching)

Output format:
CODEBLOCK22

Example 5: Special Cases Handling

Extreme Weather Warning

If extreme weather is detected (typhoon, heavy rain red warning, etc.):

CODEBLOCK23

API Call Failure

If API call fails or returns error:

CODEBLOCK24

Prompt Optimization

When actually using, can add the following prompt optimization suggestions based on situation:

• - Remind users that areas with large temperature differences between morning and evening need "onion-style" dressing
• Suggest bringing easy-to-match items to deal with weather changes
• Recommend practical layering combinations
• Consider luggage space limitations, suggest versatile items

Appendix: Script Usage Instructions

get_weather.py Script

Location: INLINECODE4

Features:

1. 1. ✅ Supports Chinese/English city names
2. ✅ Automatic relative date parsing (tomorrow, day after tomorrow, next Monday, etc.)
3. ✅ City name normalization (e.g., "Beijing" → "Beijing")
4. ✅ Returns structured JSON data
5. ✅ Includes current weather and forecast information
6. ✅ No API key required (uses wttr.in free service)

Usage:

CODEBLOCK25

Output Format:

• - stdout: JSON format weather data (for program processing)
• stderr: Human-readable formatted output (for direct viewing)

JSON Data Structure:
CODEBLOCK26

Error Handling:
If query fails, returns:
CODEBLOCK27

Dependencies:

• - Python 3.6+
• Standard library: sys, json, urllib.request, INLINECODE8
• No additional third-party libraries required

search_images.py Script

Location: INLINECODE9

Features:

1. 1. ✅ Searches fashion/outfit images related to destination
2. ✅ Supports multiple image APIs (Pexels, Pixabay, etc.)
3. ✅ Returns structured image information (URL, photographer, dimensions, etc.)
4. ✅ Automatically generates optimized search keywords
5. ✅ Provides manual search alternative

Supported APIs:

API	Requires Key	Free Quota	Recommendation
Pexels	✅ Yes	20,000/month	⭐⭐⭐⭐⭐
Pixabay

✅ Yes | 500/day | ⭐⭐⭐⭐ |
| Bing Search | ❌ No | Limited | ⭐⭐⭐ |

Usage:

CODEBLOCK28

Output Format:

• - stdout: JSON format image list (for program processing)
• stderr: Human-readable formatted output (for viewing)

JSON Data Structure:
CODEBLOCK29

Smart Keyword Generation:

Script automatically generates optimized search terms based on city context:

• - {city} street style fashion - Street style
• INLINECODE11 - Outfit inspiration
• INLINECODE12 - Casual outfits
• INLINECODE13 - Travel outfits

API Key Configuration:

Recommend using Pexels API (highest quality):

1. 1. Apply for free API key: https://www.pexels.com/api/
2. Set in script:

CODEBLOCK30

1. 3. Or via environment variable:

CODEBLOCK31

Fallback Strategy:

If API key is not configured, script will:

1. 1. Provide manual search link
2. List available API options
3. User can visit search link to view images manually

Usage Scenario:

Add image references in outfit recommendations:
CODEBLOCK32

Notes:

• - Image copyright belongs to original photographer
• For personal reference only, not for commercial use
• Respect photographers' work
• Recommend using high-quality image API for better experience