Via Platform Export Generation

This skill helps export MATSim simulation results for Via platform visualization with filtered, lightweight data.

When to Activate This Skill

Activate when user mentions:

• •"Export for Via"
• •"Generate Via visualization files"
• •"Create filtered events for Via"
• •"Prepare data for Via platform"
• •"Build agent tracks"

Workflow

Step 1: Validate Simulation Output

# Check output directory exists
ls -lh scenarios/equil/output/

# Verify required files
ls -lh scenarios/equil/output/output_plans.xml.gz
ls -lh scenarios/equil/output/output_events.xml.gz
ls -lh scenarios/equil/output/output_network.xml.gz

Step 2: Run Via Export Pipeline

Full export with all features:

python src/main/python/build_agent_tracks.py \
  --plans scenarios/equil/output/output_plans.xml.gz \
  --events scenarios/equil/output/output_events.xml.gz \
  --schedule scenarios/equil/output/output_transitSchedule.xml.gz \
  --vehicles scenarios/equil/output/output_transitVehicles.xml.gz \
  --network scenarios/equil/output/output_network.xml.gz \
  --export-filtered-events \
  --out scenarios/equil/forVia \
  --dt 5

Minimal export (events only):

python src/main/python/build_agent_tracks.py \
  --plans output/output_plans.xml.gz \
  --events output/output_events.xml.gz \
  --export-filtered-events \
  --out forVia

Step 3: Monitor Progress

Expected output with 3-checkpoint mechanism:

[1/3] Extracting agent-vehicle time ranges...
  ✓ Found 42 agent-vehicle combinations
    car_agent_01 × car_agent_01: (07:04:42-07:14:45), (14:48:40-14:58:22)
    veh_517_subway × subway: (15:36:12-15:48:54)

[2/3] Preparing event filter parameters...
  Original events: 310,743
  Agents: 50
  Vehicles: 41

[3/3] Executing fine-grained filtering...
  ✓ Filtering complete: forVia/output_events.xml

Step 4: Validate Output

# Check generated files
ls -lh forVia/

# Expected files:
# - output_events.xml (1-10 MB, filtered)
# - output_network.xml.gz (copied)
# - tracks_dt5s.csv (agent trajectories)
# - legs_table.csv (trip segments)
# - filtered_vehicles.csv (vehicle summary)
# - vehicle_usage_report.txt (statistics)

# Verify compression
wc -l forVia/output_events.xml
# Should be <<< original event count

Step 5: Check Statistics

cat forVia/vehicle_usage_report.txt

# Expected output:
# Original events: 310,743
# Filtered events: 1,212
# Compression: 99.6%
# Vehicles used: 5/2,791

Command Line Parameters

Output Files for Via

Required for Via:

1. •output_events.xml - Filtered person-vehicle events
2. •output_network.xml.gz - Network topology

Optional supporting files:

• •tracks_dt5s.csv - Agent trajectories at 5s intervals
• •filtered_vehicles.csv - Vehicle summary
• •vehicle_usage_report.txt - Statistics

Common Issues

Issue 1: Missing Python Dependencies

Error: ImportError: No module named 'lxml' Fix: pip install lxml

Issue 2: Large Unfiltered Events

Symptom: output_events.xml is same size as input Cause: Missing --export-filtered-events flag Fix: Re-run with flag

Issue 3: Empty Tracks CSV

Cause: Agent IDs mismatch or wrong time range Fix: Check agent IDs in plans vs events

Performance Benchmarks

50-agent simulation:

• •Original: 310,743 events
• •Filtered: 8,372 events (97.4% compression)
• •Vehicles: 41 used (from 2,791 total)

3-agent simulation:

• •Original: 310,743 events
• •Filtered: 1,212 events (99.6% compression)
• •Vehicles: 5 used

Validation Steps

1. •Check file sizes are reasonable (events <10MB for 50 agents)
2. •Verify compression ratio >90%
3. •Spot-check events contain only relevant agents
4. •Ensure network file is copied correctly

File References

• •Workflow: VIA_EXPORT_WORKFLOW.md
• •Python tool: src/main/python/build_agent_tracks/
• •CLAUDE.md: Lines 578-641
• •Working journal: Phase 1-4 implementation notes