DuckDuckGo weather extractor missing hourly forecast, precipitation, and icons #51

New Issue

Claude · 2026-02-15T21:11:40Z

Claude commented

2026-02-15 21:11:40 +00:00

Description

The DuckDuckGo weather extractor (sites/duckduckgo/weather.go, added in #25) returns a WeatherData struct that is missing several fields that mort's weather2 module currently extracts via inline browser scraping:

Missing Fields

Hourly forecast - mort extracts hourly temperature/condition data for a ~10-hour window. The extractor's WeatherData only has Forecast []DayForecast (daily), no hourly data.
Precipitation percentage - mort extracts the chance of precipitation (e.g., "30% chance of rain") for each day and hour. DayForecast currently only has Day, HighTemp, LowTemp, Condition - no precipitation field.
Weather icons / icon hints - mort extracts aria-label, title, and alt attributes from weather icon elements to determine the icon type (sunny, cloudy, rain, etc.). The extractor only provides a Condition string, which works but the icon attributes are more reliable since they come directly from DuckDuckGo's categorization.

Proposed Changes

type WeatherData struct {
    Location    string
    CurrentTemp float64
    Condition   string
    HighTemp    float64
    LowTemp     float64
    Humidity    string
    Wind        string
    Forecast    []DayForecast
    Hourly      []HourlyForecast  // NEW
}

type DayForecast struct {
    Day           string
    HighTemp      float64
    LowTemp       float64
    Condition     string
    Precipitation int  // NEW: percentage 0-100, -1 if unavailable
}

type HourlyForecast struct {  // NEW
    Time          string
    Temp          float64
    Condition     string
    Precipitation int    // percentage 0-100, -1 if unavailable
}

Context

Mort's pkg/logic/weather2/weather2.go has ~500 lines of inline DuckDuckGo scraping + regex parsing to extract this data. Once the extractor is complete, mort can replace all of that with a single duckduckgo.GetWeather() call. Until then, the inline scraping must remain.

Related: go-extractor #25, mort #573

## Description The DuckDuckGo weather extractor (`sites/duckduckgo/weather.go`, added in #25) returns a `WeatherData` struct that is missing several fields that mort's `weather2` module currently extracts via inline browser scraping: ### Missing Fields 1. **Hourly forecast** - mort extracts hourly temperature/condition data for a ~10-hour window. The extractor's `WeatherData` only has `Forecast []DayForecast` (daily), no hourly data. 2. **Precipitation percentage** - mort extracts the chance of precipitation (e.g., "30% chance of rain") for each day and hour. `DayForecast` currently only has `Day`, `HighTemp`, `LowTemp`, `Condition` - no precipitation field. 3. **Weather icons / icon hints** - mort extracts `aria-label`, `title`, and `alt` attributes from weather icon elements to determine the icon type (sunny, cloudy, rain, etc.). The extractor only provides a `Condition` string, which works but the icon attributes are more reliable since they come directly from DuckDuckGo's categorization. ### Proposed Changes ```go type WeatherData struct { Location string CurrentTemp float64 Condition string HighTemp float64 LowTemp float64 Humidity string Wind string Forecast []DayForecast Hourly []HourlyForecast // NEW } type DayForecast struct { Day string HighTemp float64 LowTemp float64 Condition string Precipitation int // NEW: percentage 0-100, -1 if unavailable } type HourlyForecast struct { // NEW Time string Temp float64 Condition string Precipitation int // percentage 0-100, -1 if unavailable } ``` ### Context Mort's `pkg/logic/weather2/weather2.go` has ~500 lines of inline DuckDuckGo scraping + regex parsing to extract this data. Once the extractor is complete, mort can replace all of that with a single `duckduckgo.GetWeather()` call. Until then, the inline scraping must remain. **Related:** go-extractor #25, mort #573

Claude commented

2026-02-15 21:12:40 +00:00

Starting work on this. Plan of approach:

Add HourlyForecast struct and Hourly field to WeatherData
Add Precipitation field to DayForecast
Add IconHint field to DayForecast and HourlyForecast for icon attribute extraction
Update extractWeather() to scrape hourly data, precipitation percentages, and icon hints from the weather widget
Update existing tests and add new test cases for the new fields
Update the CLI tool to display the new data

Starting work on this. Plan of approach: 1. Add `HourlyForecast` struct and `Hourly` field to `WeatherData` 2. Add `Precipitation` field to `DayForecast` 3. Add `IconHint` field to `DayForecast` and `HourlyForecast` for icon attribute extraction 4. Update `extractWeather()` to scrape hourly data, precipitation percentages, and icon hints from the weather widget 5. Update existing tests and add new test cases for the new fields 6. Update the CLI tool to display the new data

Claude referenced this issue from a commit

2026-02-15 21:22:11 +00:00

feature: add hourly forecast, precipitation, and icon hints to weather extractor

Claude referenced a pull request that will close this issue

2026-02-15 21:22:27 +00:00

feature: add hourly forecast, precipitation, and icon hints to weather extractor #52

Claude commented

2026-02-15 21:22:35 +00:00

Work finished. PR #52 adds:

HourlyForecast struct with Time, Temp, Condition, Precipitation, IconHint
Precipitation int field on DayForecast (-1 if unavailable, matching mort's convention)
IconHint string field on both DayForecast and HourlyForecast (reads aria-label > title > alt from icon elements)
extractIconHint() helper with attribute priority fallback
5 test cases covering all new fields, edge cases (no precipitation, empty doc), and icon hint priority

I cross-referenced mort's weather2 structs — the field types and -1 sentinel convention align. Mort uses Icon string on DayForecast; this PR uses IconHint since the raw value comes from DOM attributes (e.g., "PartlyCloudy") rather than a normalized icon name.

Work finished. PR #52 adds: - `HourlyForecast` struct with `Time`, `Temp`, `Condition`, `Precipitation`, `IconHint` - `Precipitation int` field on `DayForecast` (-1 if unavailable, matching mort's convention) - `IconHint string` field on both `DayForecast` and `HourlyForecast` (reads aria-label > title > alt from icon elements) - `extractIconHint()` helper with attribute priority fallback - 5 test cases covering all new fields, edge cases (no precipitation, empty doc), and icon hint priority I cross-referenced mort's `weather2` structs — the field types and -1 sentinel convention align. Mort uses `Icon string` on `DayForecast`; this PR uses `IconHint` since the raw value comes from DOM attributes (e.g., "PartlyCloudy") rather than a normalized icon name.

steve closed this issue

2026-02-15 21:23:35 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: steve/go-extractor#51