Add README

README.md

@@ -10,37 +10,74 @@ tags:
 - preference-learning
 - reward-modeling
 size_categories:
-- 10K<n<100K
+- 1K<n<10K
 ---
 
-# Web Agent Graph Dataset
+# Web Agent Grouped Graph Dataset
 
-This dataset contains web navigation tasks in graph format with single-step negatives for training reward models.
+This dataset contains web navigation tasks in grouped graph format with full history and candidate actions for training reward models.
 
 ## Dataset Description
 
-- **Format**: JSON Lines (JSONL)
-- **Size**: ~18K entries from 2.8K tasks
+- **Format**: JSON Lines (JSONL) - One entry per step with grouped candidates
+- **Size**: ~2.8K step entries from 2.8K tasks
 - **Domains**: GitLab, OpenStreetMap, Reddit, Shopping, Shopping Admin
 
 ## Data Format
 
-Each line in `graph_dataset.jsonl` is a JSON object with three main sections:
-
-### graph_data
-Task metadata including domain, root_url, and goal description.
+Each line in `graph_dataset.jsonl` represents a single step with all candidate actions grouped together:
+
+```json
+{
+  "task_id": "...",
+  "goal": "Find product X and add to cart",
+  "domain": "shopping",
+  "step_index": 3,
+  
+  "history": [
+    {"state_id": "S0", "screenshot": "...", "url": "...", "obs": "..."},
+    {"state_id": "S1", "screenshot": "...", "url": "...", "obs": "..."},
+    {"state_id": "S2", "screenshot": "...", "url": "...", "obs": "..."},
+    {"state_id": "S3", "screenshot": "...", "url": "...", "obs": "..."}
+  ],
+  
+  "current_state": {
+    "state_id": "S3",
+    "screenshot": "path/to/current.png",
+    "url": "http://...",
+    "obs": "accessibility tree..."
+  },
+  
+  "candidates": [
+    {
+      "label": "gold",
+      "action": "click('153')",
+      "next_state": "S4",
+      "next_screenshot": "path/to/next.png",
+      "next_url": "http://...",
+      "next_obs": "..."
+    },
+    {
+      "label": "negative",
+      "action": "click('88')",
+      "negative_type": "hard_negative",
+      "reason": "Leads to wrong page",
+      "next_state": "S_bad1",
+      "next_screenshot": "path/to/bad1.png",
+      "next_url": "http://...",
+      "next_obs": "..."
+    }
+  ]
+}
+```
 
-### node_data
-Current state information with URL, accessibility tree observation, and screenshot path.
+## Key Features
 
-### step_data
-Action and outcome with:
-- `action`: The action taken
-- `obs`: Next state observation
-- `step_number`: Step index
-- `is_positive`: True for gold actions, False for negatives
-- `negative_type`: Classification (easy_negative, hard_negative, detour_negative)
-- `negative_reason`: Explanation for negative actions
+- **Full History**: Complete trajectory up to current state
+- **Grouped Candidates**: All possible actions (gold + negatives) from same state
+- **Next-State Screenshots**: Screenshot paths for all candidate outcomes
+- **Rich Metadata**: Task ID, domain, goal, step index
+- **Negative Types**: Classification (easy_negative, hard_negative, detour_negative)
 
 ## Usage
 
@@ -51,9 +88,39 @@ import json
 with open('graph_dataset.jsonl', 'r') as f:
     for line in f:
         entry = json.loads(line)
-        goal = entry['graph_data']['goal']
-        action = entry['step_data']['action']
-        is_positive = entry['step_data']['is_positive']
+        
+        # Access task info
+        goal = entry['goal']
+        step_idx = entry['step_index']
+        
+        # Access history
+        history = entry['history']
+        current_state = entry['current_state']
+        
+        # Process candidates
+        for candidate in entry['candidates']:
+            if candidate['label'] == 'gold':
+                print(f"Gold action: {candidate['action']}")
+            else:
+                print(f"Negative: {candidate['action']} ({candidate['negative_type']})")
+```
+
+## Training Reward Models
+
+This format is ideal for:
+1. **Preference Learning**: Compare gold vs negative actions from same state
+2. **Reward Modeling**: Predict which action leads to goal completion
+3. **Action Ranking**: Rank all candidates by predicted reward
+
+Example:
+```python
+# For each step entry
+gold = [c for c in entry['candidates'] if c['label'] == 'gold'][0]
+negatives = [c for c in entry['candidates'] if c['label'] == 'negative']
+
+# Train model to rank gold higher than all negatives
+reward_gold = model(entry['current_state'], gold)
+rewards_neg = [model(entry['current_state'], neg) for neg in negatives]
 ```
 
 ## Statistics