vezilka-schemas 0.1.2

Pydantic models for the scraped data storage for the Vezilka Project

Vezilka Schemas

Pydantic models for managing structured scraped data in the Vezilka project.

Features

• Type-safe models using Pydantic v2
• Built-in validation for data integrity
• Easy serialization to JSON and dictionary
• Designed for MongoDB, APIs, and RAG pipelines

Package Structure

vezilka-schemas/
├── __init__.py              # Package exports
├── models.py                # Core Pydantic models

Installation

pip install vezilka-schemas

For development installation and running tests, see QUICKSTART.md

Models

Record

Represents a single scraped item.

Fields:

• id: Unique identifier
• text: Full content text
• type: Content type (RecordType)
• last_modified_at: Last modification timestamp
• meta: Metadata object (RecordMeta)

RecordMeta

Metadata associated with a record.

Fields:

• source: Source identifier (domain, filename, etc.)
• url: Original URL (optional)
• tags: List of tags (optional)
• labels: Additional labels (optional)n
• scraped_at: Scraping timestamp (optional)

RecordType

Enum defining record types:

• NARRATIVE: Articles, stories, documentation
• HUMAN: Speeches, transcripts, interviews

Usage

Creating a Record

from vezilka_schemas import Record, RecordMeta, RecordType
from datetime import datetime

# Create metadata
meta = RecordMeta(
    source="mk.wikipedia.org",
    url="https://mk.wikipedia.org/wiki/Ѓаваткол",
    tags=["Историско-географски области", "Битола"],
    labels=[],
    scraped_at=datetime.now(),
)

# Create record
record = Record(
    id="wiki_1068030",
    text="Ѓаваткол е историско-географска област...",
    type=RecordType.NARRATIVE,
    last_modified_at=datetime.now(),
    meta=meta
)

# Serialize to JSON
json_str = record.to_json()

# Convert to dictionary
data_dict = record.to_dict()

Creating from Dictionary

from vezilka_schemas import Record

data = {
    "id": "speech_9f8e7d6c5b4a",
    "text": "Говорник Африм Гаши: Почитувани пратеници...",
    "type": "human",
    "last_modified_at": "2026-02-11",
    "meta": {
        "source": "stenogram_session_12.pdf",
        "url": "",
        "tags": [],
        "labels": [],
        "scraped_at": "2026-01-17T01:40:11"
    }
}

record = Record.from_dict(data)

Validation

The models include built-in validation. For example, empty id or text will raise ValidationError.

License

MIT License

Contributing

Contributions are welcome! Please feel free to submit a pull request.