Batch Processor
SkillFiles & storageProcess multiple documents in bulk with parallel execution. Use when a user asks to batch process files, convert many documents at once, run parallel file operations, bulk rename, bulk transform, or process a directory of files concurrently. Covers parallel execution, error handling, and progress tracking.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the Batch Processor skill
What this skill tells your AI
The instructions your AI receives, as published by terminalskills/skills in skills/batch-processor/SKILL.md and read by ahel’s review.
Overview
Process multiple documents and files in bulk using parallel execution. Handles large-scale file operations including format conversion, data extraction, transformation, and validation across hundreds or thousands of files with configurable concurrency, error recovery, and progress reporting.
Instructions
When a user asks for batch processing, determine which approach fits their needs:
Task A: Parallel file processing with shell tools
For simple transformations, use xargs or GNU parallel:
# Convert all PNG files to JPEG using ImageMagick (8 parallel jobs)
find ./images -name "*.png" | xargs -P 8 -I {} bash -c \
'convert "$1" "${1%.png}.jpg"' _ {}
# Process files with GNU parallel and progress bar
find ./docs -name "*.csv" | parallel --bar --jobs 8 \
'python transform.py {} {.}_processed.csv'
# Bulk compress PDFs (4 parallel jobs)
find ./reports -name "*.pdf" | xargs -P 4 -I {} bash -c \
'gs -sDEVICE=pdfwrite -dCompatibilityLevel=1.4 -dPDFSETTINGS=/ebook \
-dNOPAUSE -dBATCH -sOutputFile="{}.compressed" "{}" && mv "{}.compressed" "{}"'
Task B: Python batch processor with concurrency control
Create a reusable batch processing script:
import asyncio
import os
from pathlib import Path
from dataclasses import dataclass, field
@dataclass
class BatchResult:
total: int = 0
success: int = 0
failed: int = 0
errors: list = field(default_factory=list)
async def process_file(filepath: Path, semaphore: asyncio.Semaphore) -> tuple[bool, str]:
async with semaphore:
try:
# Replace with actual processing logic
content = filepath.read_text()
output = content.upper() # Example transformation
out_path = filepath.with_suffix('.processed' + filepath.suffix)
out_path.write_text(output)
return True, str(filepath)
except Exception as e:
return False, f"{filepath}: {e}"
async def batch_process(
input_dir: str,
pattern: str = "*.*",
max_concurrent: int = 10
) -> BatchResult:
semaphore = asyncio.Semaphore(max_concurrent)
files = list(Path(input_dir).glob(pattern))
result = BatchResult(total=len(files))
tasks = [process_file(f, semaphore) for f in files]
for coro in asyncio.as_completed(tasks):
success, msg = await coro
if success:
result.success += 1
else:
result.failed += 1
result.errors.append(msg)
# Progress reporting
done = result.success + result.failed
print(f"\rProgress: {done}/{result.total}", end="", flush=True)
print() # Newline after progress
return result
if __name__ == "__main__":
result = asyncio.run(batch_process("./input", pattern="*.txt", max_concurrent=8))
print(f"Done: {result.success} succeeded, {result.failed} failed")
for err in result.errors:
print(f" ERROR: {err}")
Task C: Batch processing with error recovery
For long-running jobs, track progress and allow resuming:
import json
from pathlib import Path
PROGRESS_FILE = ".batch_progress.json"
def load_progress() -> set:
if Path(PROGRESS_FILE).exists():
return set(json.loads(Path(PROGRESS_FILE).read_text()))
return set()
def save_progress(completed: set):
Path(PROGRESS_FILE).write_text(json.dumps(list(completed)))
def batch_with_resume(input_dir: str, pattern: str = "*.*"):
completed = load_progress()
files = [f for f in Path(input_dir).glob(pattern) if str(f) not in completed]
print(f"Resuming: {len(completed)} done, {len(files)} remaining")
for i, filepath in enumerate(files):
try:
process_single_file(filepath) # Your processing function
completed.add(str(filepath))
if i % 10 == 0: # Checkpoint every 10 files
save_progress(completed)
except KeyboardInterrupt:
save_progress(completed)
print(f"\nSaved progress at {len(completed)} files")
raise
except Exception as e:
print(f"Error on {filepath}: {e}")
save_progress(completed)
Path(PROGRESS_FILE).unlink() # Clean up on completion
Task D: Shell-based batch with logging
#!/bin/bash
INPUT_DIR="$1"
OUTPUT_DIR="$2"
LOG_FILE="batch_$(date +%Y%m%d_%H%M%S).log"
PARALLEL_JOBS=8
TOTAL=$(find "$INPUT_DIR" -type f | wc -l)
COUNT=0
mkdir -p "$OUTPUT_DIR"
process_file() {
local file="$1"
local outfile="$OUTPUT_DIR/$(basename "$file")"
# Replace with your processing command
cp "$file" "$outfile" 2>&1
echo $?
}
export -f process_file
export OUTPUT_DIR
find "$INPUT_DIR" -type f | parallel --jobs "$PARALLEL_JOBS" --bar \
--joblog "$LOG_FILE" process_file {}
echo "Results logged to $LOG_FILE"
awk 'NR>1 {if($7!=0) fail++; else ok++} END {print ok" succeeded, "fail" failed"}' "$LOG_FILE"
Examples
Example 1: Convert a directory of Markdown files to PDF
User request: "Convert all 200 Markdown files in docs/ to PDF"
# Install pandoc if needed
# Process in parallel with 6 workers
find ./docs -name "*.md" | parallel --bar --jobs 6 \
'pandoc {} -o {.}.pdf --pdf-engine=xelatex'
echo "Conversion complete. Check for errors above."
Example 2: Extract text from hundreds of images
User request: "OCR all scanned documents in the scans/ folder"
# Using tesseract with parallel processing
find ./scans -name "*.png" -o -name "*.jpg" | parallel --bar --jobs 4 \
'tesseract {} {.} -l eng 2>/dev/null && echo "OK: {}"'
Example 3: Bulk resize images for web
User request: "Resize all product images to 800px wide, keep aspect ratio"
mkdir -p ./resized
find ./products -name "*.jpg" | xargs -P 8 -I {} bash -c \
'convert "$1" -resize 800x -quality 85 "./resized/$(basename $1)"' _ {}
echo "Resized $(ls ./resized | wc -l) images"
Guidelines
- Always test batch operations on a small subset (5-10 files) before processing the full set.
- Set a reasonable concurrency limit. Start with CPU core count for CPU-bound tasks, or 2-4x for I/O-bound tasks.
- Implement progress reporting so users can monitor long-running jobs.
- Write errors to a log file rather than stopping the entire batch.
- Create a checkpoint/resume mechanism for batches over 100 files.
- Back up original files or write output to a separate directory; never overwrite in place without confirmation.
- Use
--dry-runflags in scripts to preview operations before executing. - Monitor system resources (RAM, disk space) during large batch operations.
Signals
- GitHub stars
- 150
- Forks
- 19
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
batch-processor- Source
- github.com/terminalskills/skills