Query Cases

Query Cases — APIExample

How cases are registered

Every case is a Fragment under app/src/main/java/io/agora/api/example/examples/{basic|advanced|audio}/ with an @Example annotation:

@Example(
    index    = 10,       // unique within the group; BASIC: 0–9, ADVANCED: 10+
    group    = ADVANCED,
    name     = R.string.item_xxx,
    actionId = R.id.action_mainFragment_to_xxx,
    tipsId   = R.string.xxx_tips
)

A commented-out @Example (//@Example) means the case is disabled and won't appear in the app.

Query procedure

Step 1: Decide scope before scanning

Before listing files, ask:

• Looking for a specific API? — scan Javadoc comments for the API name; no need to read all files
• Need a free sort index? — collect all index values for the target group, then find the gap
• Listing all cases? — scan all three directories and collect annotations

Step 2: Read ARCHITECTURE.md first

Read ARCHITECTURE.md (the examples/ section of the Directory Layout). It contains a pre-built index of all cases with group, index, display name, and key API — no file scanning needed for most queries.

Use ARCHITECTURE.md as the primary source. Fall back to scanning the source directories only when:

• The query requires data not in ARCHITECTURE.md (e.g. full @Example field values, tipsId)
• ARCHITECTURE.md appears stale (a case exists in source but not in the doc)
• The output involves free-index claims, index collisions, or "is index X available?" decisions — these must be validated from source immediately before final output

Step 3: Scan case directories (fallback only)

Each .java file is a case. Subdirectories (e.g. customaudio/) contain multi-file cases — the main class is the file whose name matches the directory name (e.g. customaudio/CustomAudioSource.java). If no name match, look for the file containing @Example.

Step 4: Extract @Example fields

For each file, read the annotation for group, index, name (string resource ID), and tipsId. If the annotation is commented out, the case is disabled.

Resolve display names from app/src/main/res/values/strings.xml: R.string.item_video_snapshot → <string name="item_video_snapshot">Video Snapshot</string>

Step 5: Read class Javadoc for API mapping

The Javadoc above each class lists the key APIs demonstrated:

/**
 * This demo demonstrates how to take a snapshot of the local video stream.
 *
 * Key APIs used:
 * - RtcEngine.takeSnapshot()
 */

Use this to answer "which case uses X?" queries without reading the full implementation.

If no Javadoc is present, scan the method body for the API name as a method call. If still not found, note "API mapping unavailable" in the results table.

Step 6: Present results

Full listing — table format:

For a specific query (e.g. "which case uses takeSnapshot?"), return only matching rows.

For a free-index query, list all used indices in the target group and identify the next available slot:

BASIC range: 0–9. ADVANCED range: 10+. ADVANCED indices in use: 10, 11, 12, 15, 20 → next free: 13

Before returning any free-index/collision result, re-scan source registration points (@Example across basic/, advanced/, audio/) and recompute once from source-of-truth data.

NEVER

• NEVER count a commented-out @Example (//@Example) as an active case — it is disabled and won't appear in the app.
• NEVER mix index spaces across groups — audio/ cases use group=ADVANCED but share the same index namespace as advanced/; always scan both directories together when finding a free index.
• NEVER use filename alone to identify a subdirectory case — the main class is the file whose name matches the directory name; if no match, look for the file with @Example.
• NEVER report a free index without scanning all three directories (basic/, advanced/, audio/) for the target group — missing one causes index collisions.