Kongin

Leer en Espanol

Kongin is a Python client for OAI-PMH (Open Archives Initiative Protocol for Metadata Harvesting). It provides typed access to metadata records and easy export to DSpace REST API format.

Features

• Typed Access: Access metadata with record.title, record.creators instead of nested dicts
• Any XML Format: Handles oai_dc, xoai, oaire, simple_xml, and any other schema
• DSpace Export: Convert records to DSpace REST API JSON format (versions 7, 8, 9, 10+)
• Automatic Pagination: Iterate over all records without manual token handling
• Web Interface: Streamlit app for visual harvesting and DSpace upload
• Simple API: Clean, Pythonic interface

Installation

# Core library
pip install kongin

# With web interface
pip install kongin[web]

Usage

Kongin can be used as a Python library or through a web interface.

Option 1: Python Library (Simple)

import kongin

# One-liner harvest
records = kongin.harvest('https://repositorio.example.org/oai')
for r in records:
    print(r.title, r.creators)

# Export to DSpace JSON
kongin.export_to_dspace(records, 'output.json')

Option 2: Python Library (Full Control)

from kongin import OAIClient

# Connect to repository
client = OAIClient('https://repositorio.example.org/oai')

# Iterate over records with typed access
for record in client.list_records(metadata_prefix='oai_dc'):
    print(f"Title: {record.title}")
    print(f"Authors: {record.creators}")
    print(f"Abstract: {record.description}")

Option 3: Command Line

# Harvest and save to JSON
kongin harvest https://repositorio.example.org/oai -o records.json

# Get repository info
kongin identify https://repositorio.example.org/oai

# List available sets
kongin sets https://repositorio.example.org/oai

Option 4: Web Interface

streamlit run app.py

The web interface allows you to:

• Enter OAI-PMH URL and harvest records
• View results in a table with metrics
• Export to JSON (DSpace format) or CSV
• Upload directly to DSpace collections

Library Examples

Basic Harvesting

from kongin import OAIClient

client = OAIClient('https://repositorio.example.org/oai')

# List available sets
sets = client.list_sets()
for s in sets:
    print(f"{s['set_spec']}: {s['set_name']}")

# List metadata formats
formats = client.list_metadata_formats()
for f in formats:
    print(f"{f['prefix']}: {f['namespace']}")

# Get repository info
info = client.identify()
print(f"Repository: {info['repository_name']}")

Harvesting Records

# Harvest all records from a set
for record in client.list_records(
    metadata_prefix='simple_xml',
    set_spec='articles'
):
    print(record.title)
    print(record.creators)

# With date filtering
for record in client.list_records(
    metadata_prefix='oai_dc',
    from_date='2024-01-01',
    until_date='2024-12-31'
):
    process(record)

# Get a single record
record = client.get_record(
    identifier='oai:repo.example.org:12345',
    metadata_prefix='oai_dc'
)

Accessing Metadata

# Typed properties for common fields
print(record.title)        # First title
print(record.titles)       # All titles
print(record.creators)     # All authors
print(record.description)  # First description/abstract
print(record.date)         # First date
print(record.subjects)     # All subjects
print(record.identifiers)  # All identifiers (DOI, URI, etc.)

# Access any field directly
volume = record.metadata.get('oaire:citationVolume')
issue = record.metadata.get('oaire:citationIssue')
custom = record.metadata.get('custom:field')

# Get all values for a field
all_rights = record.metadata.get_all('dc:rights')

# Check if field exists
if 'dcterms:abstract' in record.metadata:
    print(record.metadata.get('dcterms:abstract'))

Export to DSpace REST API

DSpace changed to a REST API starting from version 7. Kongin supports DSpace 7, 8, 9, 10 and future versions that maintain API compatibility.

from kongin import OAIClient, DSpaceExporter

client = OAIClient('https://repositorio.example.org/oai')

# Export single record
record = client.get_record('oai:repo:123', 'oai_dc')
dspace_item = record.to_dspace()
# Result is JSON compatible with POST /api/core/items

# Export multiple records
records = list(client.list_records(set_spec='theses'))
exporter = DSpaceExporter()
items = exporter.export_records(records)

# Save to JSON file
exporter.save_json(records, 'dspace_import.json')

# Custom field mapping
custom_mapping = {
    'local:category': 'local.category',
    'oaire:citationVolume': 'local.citation.volume',
}
exporter = DSpaceExporter(custom_mapping)
items = exporter.export_records(records)

Upload to DSpace

from kongin import OAIClient, DSpaceClient

# Harvest records
oai = OAIClient('https://source-repo.org/oai')
records = list(oai.list_records(metadata_prefix='oai_dc', set_spec='articles'))

# Upload to DSpace
dspace = DSpaceClient(
    base_url='https://dspace.example.org',
    email='admin@example.org',
    password='password'
)

# List collections
collections = dspace.list_collections()
for c in collections:
    print(f"{c['name']} ({c['uuid']})")

# Upload to a collection
collection_id = collections[0]['uuid']
for record in records:
    dspace.create_item(record, collection_id)

Manual Pagination

# Get first page
page = client.list_records_page(metadata_prefix='oai_dc')
print(f"Total records: {page.complete_list_size}")

for record in page:
    process(record)

# Get next pages
while page.has_more:
    page = client.resume(page.resumption_token)
    for record in page:
        process(record)

Configuration

client = OAIClient(
    url='https://repositorio.example.org/oai',
    timeout=30,           # Request timeout in seconds
    max_retries=3,        # Retry failed requests
    http_method='GET',    # or 'POST'
    requests_args={       # Passed to requests library
        'verify': False,  # Disable SSL verification
        'headers': {'User-Agent': 'MyHarvester/1.0'}
    }
)

API Reference

OAIClient

• identify() - Get repository information
• list_sets() - List available sets
• list_metadata_formats() - List supported formats
• get_record(identifier, metadata_prefix) - Get single record
• list_records(...) - Iterate all records (auto-pagination)
• list_records_page(...) - Get single page (manual pagination)
• list_identifiers(...) - Iterate record headers only
• harvest(**params) - Raw OAI-PMH request

Record

• .identifier - OAI identifier
• .datestamp - Last modified date
• .set_specs - Sets this record belongs to
• .deleted - True if record was deleted
• .metadata - Metadata container
• .title, .creators, .description, etc. - Typed properties
• .to_dict() - Export as dictionary
• .to_dspace() - Export for DSpace REST API

Metadata

• .get(key) - Get first value
• .get_all(key) - Get all values as list
• .to_dict() - Export as dictionary

DSpaceExporter

• .export_record(record) - Convert single record
• .export_records(records) - Convert multiple records
• .to_json(records) - Export as JSON string
• .save_json(records, filepath) - Save to file

DSpaceClient

• .list_collections() - Get available collections
• .list_communities() - Get available communities
• .create_item(record, collection_id) - Create item in collection
• .upload_records(records, collection_id) - Upload multiple records

DSpace Compatibility

This library is compatible with DSpace REST API versions:

DSpace Version	API	Status
DSpace 7.x	REST API v7	Supported
DSpace 8.x	REST API v7	Supported
DSpace 9.x	REST API v7	Supported
DSpace 10.x	REST API v7	Supported

Note: DSpace versions prior to 7 used a different API (XMLUI/JSPUI) and are not supported.

License

MIT License - see LICENSE file.

Author

Haroldo Vivallo

Acknowledgments

This project was developed with assistance from Claude (Anthropic). The code was reviewed, tested, and validated by the author. Kongin builds upon an earlier prototype to create a more robust and feature-complete OAI-PMH client.