RSS Parser

About

rss-parser is a type-safe Python RSS/Atom parsing module built using pydantic and xmltodict.

Installation

pip install rss-parser

or

git clone https://github.com/dhvcc/rss-parser.git
cd rss-parser
poetry build
pip install dist/*.whl

V1 -> V2 Migration

• The Parser class has been renamed to RSSParser
• Models for RSS-specific schemas have been moved from rss_parser.models to rss_parser.models.rss. Generic types remain unchanged
• Date parsing has been improved and now uses pydantic's validator instead of email.utils, producing better datetime objects where it previously defaulted to str

V2 -> V3 Migration

rss-parser 3.x upgrades the runtime models to Pydantic v2. Highlights:

• New default models now inherit from pydantic.BaseModel v2 and use model_validate/model_dump. If you extend our classes, switch from dict()/json() to model_dump()/model_dump_json().
• Legacy compatibility lives under rss_parser.models.legacy. Point your custom parser at the legacy schema if you must stay on the v1 API surface.
• Collections: list-like XML fields now use OnlyList[...] directly with an automatic default_factory so that attributes are always lists (no more Optional[OnlyList[T]] = Field(..., default=[])). Update custom schemas accordingly.
• Custom hooks: if you relied on rss_parser.pydantic_proxy, import it from rss_parser.models.legacy.pydantic_proxy. The top-level module only re-exports it for backwards compatibility.

See the “Legacy Models” section below for sample snippets showing how to stay on the older types. Tests in this repo cover both tracks to guarantee matching output.

Legacy Models

Pydantic v1-based models are still available under rss_parser.models.legacy. They retain the previous behaviour and re-export the import_v1_pydantic helper as rss_parser.models.legacy.pydantic_proxy.import_v1_pydantic. You can continue to use them by pointing your parser at the legacy schema:

from rss_parser import RSSParser
from rss_parser.models.legacy.rss import RSS as LegacyRSS

class LegacyRSSParser(RSSParser):
    schema = LegacyRSS

Tests in this repository run against both the v2 and legacy models to ensure parity.

Usage

Quickstart

NOTE: For parsing Atom, use AtomParser

from rss_parser import RSSParser
from requests import get  # noqa

rss_url = "https://rss.art19.com/apology-line"
response = get(rss_url)

rss = RSSParser.parse(response.text)

# Print out rss meta data
print("Language", rss.channel.language)
print("RSS", rss.version)

# Iteratively print feed items
for item in rss.channel.items:
    print(item.title)
    print(item.description[:50])

# Language en
# RSS 2.0
# Wondery Presents - Flipping The Bird: Elon vs Twitter
# <p>When Elon Musk posted a video of himself arrivi
# Introducing: The Apology Line
# <p>If you could call a number and say you’re sorry

Here we can see that the description still contains <p> tags - this is because it's wrapped in CDATA like so:

<![CDATA[<p>If you could call ...</p>]]>

Overriding Schema

If you want to customize the schema or provide a custom one, use the schema keyword argument of the parser:

from rss_parser import RSSParser
from rss_parser.models import XMLBaseModel
from rss_parser.models.rss import RSS
from rss_parser.models.types import Tag


class CustomSchema(RSS, XMLBaseModel):
    channel: None = None  # Removing previous channel field
    custom: Tag[str]


with open("tests/samples/custom.xml") as f:
    data = f.read()

rss = RSSParser.parse(data, schema=CustomSchema)

print("RSS", rss.version)
print("Custom", rss.custom)

# RSS 2.0
# Custom Custom tag data

xmltodict

This library uses xmltodict to parse XML data. You can find the detailed documentation here.

The key thing to understand is that your data is processed into dictionaries.

For example, this XML:

<tag>content</tag>

will result in the following dictionary:

{
    "tag": "content"
}

However, when handling attributes, the content of the tag will also be a dictionary:

<tag attr="1" data-value="data">data</tag>

This becomes:

{
    "tag": {
        "@attr": "1",
        "@data-value": "data",
        "#text": "content"
    }
}

Multiple children of a tag will be placed into a list:

<div>
    <tag>content</tag>
    <tag>content2</tag>
</div>

This results in a list:

[
    { "tag": "content" },
    { "tag": "content" },
]

If you don't want to deal with these conditions and want to parse something always as a list, please use rss_parser.models.types.only_list.OnlyList like we did in Channel:

from typing import Optional

from pydantic import Field

from rss_parser.models.rss.item import Item
from rss_parser.models.types.only_list import OnlyList
from rss_parser.models.types.tag import Tag
...


class OptionalChannelElementsMixin(...):
    ...
    items: Optional[OnlyList[Tag[Item]]] = Field(alias="item", default_factory=list)

Tag Field

This is a generic field that handles tags as raw data or as a dictionary returned with attributes.

Example:

from rss_parser.models import XMLBaseModel
from rss_parser.models.types.tag import Tag


class Model(XMLBaseModel):
    width: Tag[int]
    category: Tag[str]


m = Model(
    width=48,
    category={"@someAttribute": "https://example.com", "#text": "valid string"},
)

# Content value is an integer, as per the generic type
assert m.width.content == 48

assert type(m.width), type(m.width.content) == (Tag[int], int)

# The attributes are empty by default
assert m.width.attributes == {} # But are populated when provided.

# Note that the @ symbol is trimmed from the beginning and the name is converted to snake_case
assert m.category.attributes == {'some_attribute': 'https://example.com'}

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Install dependencies with poetry install (pip install poetry).

Using pre-commit is highly recommended. To install hooks, run:

poetry run pre-commit install -t=pre-commit -t=pre-push

License

GPLv3