fasthtmlparser 0.1.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- fasthtmlparser-0.1.0/.gitattributes +9 -0
- fasthtmlparser-0.1.0/.github/workflows/ci.yml +117 -0
- fasthtmlparser-0.1.0/.gitignore +6 -0
- fasthtmlparser-0.1.0/Cargo.lock +148 -0
- fasthtmlparser-0.1.0/Cargo.toml +18 -0
- fasthtmlparser-0.1.0/LICENSE.md +79 -0
- fasthtmlparser-0.1.0/PKG-INFO +156 -0
- fasthtmlparser-0.1.0/README.md +131 -0
- fasthtmlparser-0.1.0/bench-results.md +28 -0
- fasthtmlparser-0.1.0/pyproject.toml +40 -0
- fasthtmlparser-0.1.0/python/fasthtmlparser/__init__.py +38 -0
- fasthtmlparser-0.1.0/python/fasthtmlparser/_fasthtmlparser.pyi +342 -0
- fasthtmlparser-0.1.0/python/fasthtmlparser/py.typed +0 -0
- fasthtmlparser-0.1.0/src/lib.rs +867 -0
- fasthtmlparser-0.1.0/tests/test_api.py +143 -0
- fasthtmlparser-0.1.0/tests/test_conformance.py +99 -0
|
@@ -0,0 +1,9 @@
|
|
|
1
|
+
# Sources with LF everywhere.
|
|
2
|
+
*.rs text eol=lf
|
|
3
|
+
*.toml text eol=lf
|
|
4
|
+
*.py text eol=lf
|
|
5
|
+
*.pyi text eol=lf
|
|
6
|
+
|
|
7
|
+
# Test pages and the expected .NET output are compared byte for byte (positions are byte offsets):
|
|
8
|
+
# git must never convert their line endings.
|
|
9
|
+
testdata/** -text linguist-vendored
|
|
@@ -0,0 +1,117 @@
|
|
|
1
|
+
name: CI
|
|
2
|
+
|
|
3
|
+
on:
|
|
4
|
+
push:
|
|
5
|
+
branches: [main]
|
|
6
|
+
tags: ["v*"]
|
|
7
|
+
pull_request:
|
|
8
|
+
|
|
9
|
+
permissions:
|
|
10
|
+
contents: read
|
|
11
|
+
|
|
12
|
+
jobs:
|
|
13
|
+
test:
|
|
14
|
+
runs-on: ${{ matrix.os }}
|
|
15
|
+
strategy:
|
|
16
|
+
matrix:
|
|
17
|
+
os: [ubuntu-latest, windows-latest, macos-latest]
|
|
18
|
+
python: ["3.10", "3.14"]
|
|
19
|
+
steps:
|
|
20
|
+
- uses: actions/checkout@v4
|
|
21
|
+
- uses: actions/setup-python@v5
|
|
22
|
+
with:
|
|
23
|
+
python-version: ${{ matrix.python }}
|
|
24
|
+
- uses: dtolnay/rust-toolchain@stable
|
|
25
|
+
- name: Build and test
|
|
26
|
+
shell: bash
|
|
27
|
+
run: |
|
|
28
|
+
python -m venv .venv
|
|
29
|
+
if [ -f .venv/bin/activate ]; then . .venv/bin/activate; else . .venv/Scripts/activate; fi
|
|
30
|
+
pip install maturin pytest
|
|
31
|
+
maturin develop --release
|
|
32
|
+
pytest -q
|
|
33
|
+
|
|
34
|
+
lint:
|
|
35
|
+
runs-on: ubuntu-latest
|
|
36
|
+
steps:
|
|
37
|
+
- uses: actions/checkout@v4
|
|
38
|
+
- uses: dtolnay/rust-toolchain@stable
|
|
39
|
+
with:
|
|
40
|
+
components: clippy, rustfmt
|
|
41
|
+
- run: cargo fmt --check
|
|
42
|
+
- run: cargo clippy -- -D warnings
|
|
43
|
+
|
|
44
|
+
# Every library of the benchmark must extract the same records.
|
|
45
|
+
bench:
|
|
46
|
+
runs-on: ubuntu-latest
|
|
47
|
+
steps:
|
|
48
|
+
- uses: actions/checkout@v4
|
|
49
|
+
- uses: actions/setup-python@v5
|
|
50
|
+
with:
|
|
51
|
+
python-version: "3.14"
|
|
52
|
+
- uses: dtolnay/rust-toolchain@stable
|
|
53
|
+
- run: |
|
|
54
|
+
python -m venv .venv && . .venv/bin/activate
|
|
55
|
+
pip install maturin selectolax lxml parsel beautifulsoup4
|
|
56
|
+
maturin develop --release
|
|
57
|
+
python bench/bench.py --check
|
|
58
|
+
|
|
59
|
+
# abi3 wheels (one per platform for Python 3.10+) and the sdist.
|
|
60
|
+
wheels:
|
|
61
|
+
needs: [test, lint]
|
|
62
|
+
runs-on: ${{ matrix.os }}
|
|
63
|
+
strategy:
|
|
64
|
+
matrix:
|
|
65
|
+
include:
|
|
66
|
+
- { os: ubuntu-latest, target: x86_64, manylinux: auto }
|
|
67
|
+
- { os: ubuntu-latest, target: aarch64, manylinux: auto }
|
|
68
|
+
- { os: ubuntu-latest, target: x86_64, manylinux: musllinux_1_2 }
|
|
69
|
+
- { os: ubuntu-latest, target: aarch64, manylinux: musllinux_1_2 }
|
|
70
|
+
- { os: windows-latest, target: x64 }
|
|
71
|
+
- { os: windows-11-arm, target: aarch64 }
|
|
72
|
+
- { os: macos-latest, target: aarch64 }
|
|
73
|
+
- { os: macos-15-intel, target: x86_64 }
|
|
74
|
+
steps:
|
|
75
|
+
- uses: actions/checkout@v4
|
|
76
|
+
- uses: actions/setup-python@v5
|
|
77
|
+
with:
|
|
78
|
+
python-version: "3.14"
|
|
79
|
+
- uses: PyO3/maturin-action@v1
|
|
80
|
+
with:
|
|
81
|
+
target: ${{ matrix.target }}
|
|
82
|
+
manylinux: ${{ matrix.manylinux }}
|
|
83
|
+
args: --release --out dist
|
|
84
|
+
- uses: actions/upload-artifact@v4
|
|
85
|
+
with:
|
|
86
|
+
name: wheels-${{ matrix.os }}-${{ matrix.target }}-${{ matrix.manylinux }}
|
|
87
|
+
path: dist
|
|
88
|
+
|
|
89
|
+
sdist:
|
|
90
|
+
needs: [test, lint]
|
|
91
|
+
runs-on: ubuntu-latest
|
|
92
|
+
steps:
|
|
93
|
+
- uses: actions/checkout@v4
|
|
94
|
+
- uses: PyO3/maturin-action@v1
|
|
95
|
+
with:
|
|
96
|
+
command: sdist
|
|
97
|
+
args: --out dist
|
|
98
|
+
- uses: actions/upload-artifact@v4
|
|
99
|
+
with:
|
|
100
|
+
name: sdist
|
|
101
|
+
path: dist
|
|
102
|
+
|
|
103
|
+
# A version tag (v0.1.0) publishes to PyPI with trusted publishing: no token is stored in the repository.
|
|
104
|
+
# PyPI must list this repository, this workflow (ci.yml) and the environment "pypi" as a trusted publisher.
|
|
105
|
+
publish:
|
|
106
|
+
if: startsWith(github.ref, 'refs/tags/v')
|
|
107
|
+
needs: [wheels, sdist, bench]
|
|
108
|
+
runs-on: ubuntu-latest
|
|
109
|
+
environment: pypi
|
|
110
|
+
permissions:
|
|
111
|
+
id-token: write
|
|
112
|
+
steps:
|
|
113
|
+
- uses: actions/download-artifact@v4
|
|
114
|
+
with:
|
|
115
|
+
path: dist
|
|
116
|
+
merge-multiple: true
|
|
117
|
+
- uses: pypa/gh-action-pypi-publish@release/v1
|
|
@@ -0,0 +1,148 @@
|
|
|
1
|
+
# This file is automatically @generated by Cargo.
|
|
2
|
+
# It is not intended for manual editing.
|
|
3
|
+
version = 4
|
|
4
|
+
|
|
5
|
+
[[package]]
|
|
6
|
+
name = "fasthtmlparser"
|
|
7
|
+
version = "0.1.0"
|
|
8
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
9
|
+
checksum = "7d35fd1e20a7a369d46059753584f38a6605ce7af65edab5babd6251ad2dbfed"
|
|
10
|
+
dependencies = [
|
|
11
|
+
"memchr",
|
|
12
|
+
]
|
|
13
|
+
|
|
14
|
+
[[package]]
|
|
15
|
+
name = "fasthtmlparser-py"
|
|
16
|
+
version = "0.1.0"
|
|
17
|
+
dependencies = [
|
|
18
|
+
"fasthtmlparser",
|
|
19
|
+
"pyo3",
|
|
20
|
+
]
|
|
21
|
+
|
|
22
|
+
[[package]]
|
|
23
|
+
name = "heck"
|
|
24
|
+
version = "0.5.0"
|
|
25
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
26
|
+
checksum = "2304e00983f87ffb38b55b444b5e3b60a884b5d30c0fca7d82fe33449bbe55ea"
|
|
27
|
+
|
|
28
|
+
[[package]]
|
|
29
|
+
name = "libc"
|
|
30
|
+
version = "0.2.191"
|
|
31
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
32
|
+
checksum = "65cd87c9a6d60536580b316ad985726104730a5151c99845f2490428ef91240a"
|
|
33
|
+
|
|
34
|
+
[[package]]
|
|
35
|
+
name = "memchr"
|
|
36
|
+
version = "2.8.3"
|
|
37
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
38
|
+
checksum = "cf8baf1c55e62ffcace7a9f06f4bd9cd3f0c4beb022d3b367256b91b87513d98"
|
|
39
|
+
|
|
40
|
+
[[package]]
|
|
41
|
+
name = "once_cell"
|
|
42
|
+
version = "1.21.4"
|
|
43
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
44
|
+
checksum = "9f7c3e4beb33f85d45ae3e3a1792185706c8e16d043238c593331cc7cd313b50"
|
|
45
|
+
|
|
46
|
+
[[package]]
|
|
47
|
+
name = "portable-atomic"
|
|
48
|
+
version = "1.15.0"
|
|
49
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
50
|
+
checksum = "05c8b63e8d9609db387f0324918f81d68fe27748f084ef092fb35954d0539a85"
|
|
51
|
+
|
|
52
|
+
[[package]]
|
|
53
|
+
name = "proc-macro2"
|
|
54
|
+
version = "1.0.107"
|
|
55
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
56
|
+
checksum = "985e7ec9bb745e6ce6535b544d84d6cd6f7ad8bd711c398938ae983b91a766d9"
|
|
57
|
+
dependencies = [
|
|
58
|
+
"unicode-ident",
|
|
59
|
+
]
|
|
60
|
+
|
|
61
|
+
[[package]]
|
|
62
|
+
name = "pyo3"
|
|
63
|
+
version = "0.30.1"
|
|
64
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
65
|
+
checksum = "229ca13b8a15bb0d1487b81ab975bfcd97513416499a8312011b6036e42e0de4"
|
|
66
|
+
dependencies = [
|
|
67
|
+
"libc",
|
|
68
|
+
"once_cell",
|
|
69
|
+
"portable-atomic",
|
|
70
|
+
"pyo3-build-config",
|
|
71
|
+
"pyo3-ffi",
|
|
72
|
+
"pyo3-macros",
|
|
73
|
+
]
|
|
74
|
+
|
|
75
|
+
[[package]]
|
|
76
|
+
name = "pyo3-build-config"
|
|
77
|
+
version = "0.30.1"
|
|
78
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
79
|
+
checksum = "70130e2b6dd1ee589a8788026a429561ca5c40fa7a2fa7914c2282eddae24700"
|
|
80
|
+
dependencies = [
|
|
81
|
+
"target-lexicon",
|
|
82
|
+
]
|
|
83
|
+
|
|
84
|
+
[[package]]
|
|
85
|
+
name = "pyo3-ffi"
|
|
86
|
+
version = "0.30.1"
|
|
87
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
88
|
+
checksum = "5e7ff73c555c286d212d4ce7e5fbda845a973c4f61bd5f160f5f2c3c2a04d1c4"
|
|
89
|
+
dependencies = [
|
|
90
|
+
"libc",
|
|
91
|
+
"pyo3-build-config",
|
|
92
|
+
]
|
|
93
|
+
|
|
94
|
+
[[package]]
|
|
95
|
+
name = "pyo3-macros"
|
|
96
|
+
version = "0.30.1"
|
|
97
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
98
|
+
checksum = "c46d2e5862c7bb8207e3b639f713dda3c499933444784154c11d013c5a691b60"
|
|
99
|
+
dependencies = [
|
|
100
|
+
"proc-macro2",
|
|
101
|
+
"pyo3-macros-backend",
|
|
102
|
+
"quote",
|
|
103
|
+
"syn",
|
|
104
|
+
]
|
|
105
|
+
|
|
106
|
+
[[package]]
|
|
107
|
+
name = "pyo3-macros-backend"
|
|
108
|
+
version = "0.30.1"
|
|
109
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
110
|
+
checksum = "38b89f066f2ae72a2aef52d1537ca2fcd104440ebfadee06e97aa90738a25b9b"
|
|
111
|
+
dependencies = [
|
|
112
|
+
"heck",
|
|
113
|
+
"proc-macro2",
|
|
114
|
+
"quote",
|
|
115
|
+
"syn",
|
|
116
|
+
]
|
|
117
|
+
|
|
118
|
+
[[package]]
|
|
119
|
+
name = "quote"
|
|
120
|
+
version = "1.0.47"
|
|
121
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
122
|
+
checksum = "1fbf4db142a473a8d80c26bbf18454ed458bf8d26c8219c331daecfdbd079001"
|
|
123
|
+
dependencies = [
|
|
124
|
+
"proc-macro2",
|
|
125
|
+
]
|
|
126
|
+
|
|
127
|
+
[[package]]
|
|
128
|
+
name = "syn"
|
|
129
|
+
version = "3.0.10"
|
|
130
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
131
|
+
checksum = "4ae9ccbea3a7b0b09443036490c8f34a8b1b760fd345e5c0dae0aac718faea79"
|
|
132
|
+
dependencies = [
|
|
133
|
+
"proc-macro2",
|
|
134
|
+
"quote",
|
|
135
|
+
"unicode-ident",
|
|
136
|
+
]
|
|
137
|
+
|
|
138
|
+
[[package]]
|
|
139
|
+
name = "target-lexicon"
|
|
140
|
+
version = "0.13.6"
|
|
141
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
142
|
+
checksum = "8c447be3a2abfa84b2d91680c56600bd6b2873866de27b36ec911be705e291da"
|
|
143
|
+
|
|
144
|
+
[[package]]
|
|
145
|
+
name = "unicode-ident"
|
|
146
|
+
version = "1.0.27"
|
|
147
|
+
source = "registry+https://github.com/rust-lang/crates.io-index"
|
|
148
|
+
checksum = "a2c754d6c33795a1c324727428e5a7dedb5b06195f9890bdbcba760d3e246563"
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
[package]
|
|
2
|
+
name = "fasthtmlparser-py"
|
|
3
|
+
version = "0.1.0"
|
|
4
|
+
edition = "2021"
|
|
5
|
+
publish = false
|
|
6
|
+
readme = "README.md"
|
|
7
|
+
|
|
8
|
+
[lib]
|
|
9
|
+
name = "_fasthtmlparser"
|
|
10
|
+
crate-type = ["cdylib"]
|
|
11
|
+
|
|
12
|
+
[dependencies]
|
|
13
|
+
fasthtmlparser = "0.1"
|
|
14
|
+
pyo3 = { version = "0.30", features = ["extension-module", "abi3-py310", "generate-import-lib"] }
|
|
15
|
+
|
|
16
|
+
[profile.release]
|
|
17
|
+
lto = "fat"
|
|
18
|
+
codegen-units = 1
|
|
@@ -0,0 +1,79 @@
|
|
|
1
|
+
# z3l.FastHtmlParser License
|
|
2
|
+
|
|
3
|
+
Copyright © 2026 z3l. All rights reserved.
|
|
4
|
+
|
|
5
|
+
This license applies to the z3l.FastHtmlParser library, its source code, binaries and
|
|
6
|
+
NuGet packages (the "Software").
|
|
7
|
+
|
|
8
|
+
## 1. Definitions
|
|
9
|
+
|
|
10
|
+
- **"Licensee"** — the individual or legal entity that uses the Software, together with its
|
|
11
|
+
Affiliates.
|
|
12
|
+
- **"Affiliate"** — any entity that controls, is controlled by, or is under common control with
|
|
13
|
+
the Licensee.
|
|
14
|
+
- **"Annual Revenue"** — the total gross revenue of the Licensee (including all Affiliates) for
|
|
15
|
+
the last completed fiscal year, in US dollars or equivalent.
|
|
16
|
+
- **"Headcount"** — the total number of employees and individual contractors working for the
|
|
17
|
+
Licensee (including all Affiliates) on a full-time or full-time-equivalent basis, at any time
|
|
18
|
+
during the last 12 months.
|
|
19
|
+
- **"Small Licensee"** — a Licensee whose Annual Revenue is **below USD 1,000,000** and whose
|
|
20
|
+
Headcount is **below 100**. Both conditions must be met.
|
|
21
|
+
|
|
22
|
+
## 2. Free License (Small Licensees)
|
|
23
|
+
|
|
24
|
+
If the Licensee is a Small Licensee (Annual Revenue **below USD 1,000,000** and Headcount
|
|
25
|
+
**below 100**), the Licensee is granted, free of charge, a worldwide, non-exclusive,
|
|
26
|
+
non-transferable license to use, copy and modify the Software and to distribute it as part of the
|
|
27
|
+
Licensee's own applications and services, including commercial ones.
|
|
28
|
+
|
|
29
|
+
## 3. Commercial License (all other Licensees)
|
|
30
|
+
|
|
31
|
+
If the Licensee's Annual Revenue is **USD 1,000,000 or more**, or its Headcount is **100 or
|
|
32
|
+
more**, the Licensee must purchase a Commercial License.
|
|
33
|
+
|
|
34
|
+
- Price: a **one-time payment of USD 1,000** per Licensee (the Licensee together with all its
|
|
35
|
+
Affiliates).
|
|
36
|
+
- The Commercial License is perpetual and covers all current and future versions of the
|
|
37
|
+
Software, any number of developers, applications, servers and deployments of the Licensee.
|
|
38
|
+
- A Licensee that stops being a Small Licensee (its Annual Revenue or Headcount reaches the
|
|
39
|
+
threshold) while already using the Software must purchase the Commercial License within
|
|
40
|
+
30 days.
|
|
41
|
+
- To purchase: https://romanti3172.github.io/z3l.FastHtmlParser/#pricing, or contact **romanti3172@gmail.com** for an invoice.
|
|
42
|
+
|
|
43
|
+
The Commercial License grants the same rights as section 2, without the revenue and headcount
|
|
44
|
+
limits.
|
|
45
|
+
|
|
46
|
+
**One license for all z3l libraries.** The Commercial License is not limited to the Software: it
|
|
47
|
+
also covers z3l.Scraper and every other library and package published by z3l, now or in the
|
|
48
|
+
future, on the same terms. The Licensee pays USD 1,000 once and needs no other z3l license.
|
|
49
|
+
|
|
50
|
+
## 4. Conditions (both licenses)
|
|
51
|
+
|
|
52
|
+
1. This license text and the copyright notice must be kept in all copies or substantial portions
|
|
53
|
+
of the Software's source code.
|
|
54
|
+
2. The Licensee may not sell, sublicense or redistribute the Software on its own, or as part of a
|
|
55
|
+
library, SDK or component whose main purpose is HTML/XML parsing, so that it competes with the
|
|
56
|
+
Software.
|
|
57
|
+
3. Modified versions of the Software may not be published or distributed under a different
|
|
58
|
+
license.
|
|
59
|
+
4. End users of the Licensee's applications and services do not need their own license to use
|
|
60
|
+
those applications and services.
|
|
61
|
+
|
|
62
|
+
## 5. Termination
|
|
63
|
+
|
|
64
|
+
The license terminates automatically if the Licensee breaches it and does not cure the breach
|
|
65
|
+
within 30 days after becoming aware of it. On termination the Licensee must stop using and
|
|
66
|
+
distributing the Software.
|
|
67
|
+
|
|
68
|
+
## 6. No Warranty
|
|
69
|
+
|
|
70
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT
|
|
71
|
+
NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
|
72
|
+
NONINFRINGEMENT.
|
|
73
|
+
|
|
74
|
+
## 7. Limitation of Liability
|
|
75
|
+
|
|
76
|
+
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
77
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
|
|
78
|
+
CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. THE TOTAL LIABILITY OF
|
|
79
|
+
THE COPYRIGHT HOLDERS SHALL NOT EXCEED THE AMOUNT PAID BY THE LICENSEE FOR THE SOFTWARE.
|
|
@@ -0,0 +1,156 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: fasthtmlparser
|
|
3
|
+
Version: 0.1.0
|
|
4
|
+
Classifier: Development Status :: 4 - Beta
|
|
5
|
+
Classifier: Intended Audience :: Developers
|
|
6
|
+
Classifier: License :: Other/Proprietary License
|
|
7
|
+
Classifier: Programming Language :: Python :: 3
|
|
8
|
+
Classifier: Programming Language :: Python :: 3 :: Only
|
|
9
|
+
Classifier: Programming Language :: Rust
|
|
10
|
+
Classifier: Topic :: Internet :: WWW/HTTP :: Indexing/Search
|
|
11
|
+
Classifier: Topic :: Text Processing :: Markup :: HTML
|
|
12
|
+
Classifier: Topic :: Text Processing :: Markup :: XML
|
|
13
|
+
Classifier: Typing :: Typed
|
|
14
|
+
License-File: LICENSE.md
|
|
15
|
+
Summary: Fast HTML / XML parser for scraping with XPath 1.0 and CSS selectors, in Rust. A port of z3l.FastHtmlParser.
|
|
16
|
+
Keywords: html,parser,xpath,css,selectors,scraping,web-scraping,rust
|
|
17
|
+
Author: z3l
|
|
18
|
+
Requires-Python: >=3.10
|
|
19
|
+
Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
|
|
20
|
+
Project-URL: Homepage, https://github.com/romanti3172/fasthtmlparser-py
|
|
21
|
+
Project-URL: Issues, https://github.com/romanti3172/fasthtmlparser-py/issues
|
|
22
|
+
Project-URL: Rust crate, https://crates.io/crates/fasthtmlparser
|
|
23
|
+
Project-URL: Source, https://github.com/romanti3172/fasthtmlparser-py
|
|
24
|
+
|
|
25
|
+
# fasthtmlparser
|
|
26
|
+
|
|
27
|
+
[](https://github.com/romanti3172/fasthtmlparser-py/actions/workflows/ci.yml)
|
|
28
|
+
[](https://pypi.org/project/fasthtmlparser/)
|
|
29
|
+
|
|
30
|
+
A fast HTML / XML parser for scraping, with **XPath 1.0** and **CSS selectors**: Python bindings of the
|
|
31
|
+
[Rust port](https://github.com/romanti3172/fasthtmlparser-rs) of
|
|
32
|
+
[z3l.FastHtmlParser](https://github.com/romanti3172/z3l.FastHtmlParser). It is built for extracting data, not for
|
|
33
|
+
building a DOM: a page becomes a flat list of tags that hold only positions in the source; text, attributes and
|
|
34
|
+
entities are read when you ask for them. Parsing is **5–7x faster than selectolax** and **20x faster than lxml**
|
|
35
|
+
(see below).
|
|
36
|
+
|
|
37
|
+
```
|
|
38
|
+
pip install fasthtmlparser
|
|
39
|
+
```
|
|
40
|
+
|
|
41
|
+
Python 3.10 or later. Wheels for Linux (glibc and musl), Windows and macOS, x86-64 and ARM64; no dependencies.
|
|
42
|
+
|
|
43
|
+
```python
|
|
44
|
+
import fasthtmlparser as fh
|
|
45
|
+
|
|
46
|
+
doc = fh.parse(html, url="https://books.toscrape.com/") # str or bytes; url resolves relative links
|
|
47
|
+
|
|
48
|
+
# CSS selectors, as querySelectorAll / querySelector
|
|
49
|
+
for pod in doc.css("article.product_pod"):
|
|
50
|
+
title = pod.css_first("h3 a").attr("title")
|
|
51
|
+
price = pod.css_first("p.price_color").text
|
|
52
|
+
link = pod.css_first("h3 a").href # absolute URL
|
|
53
|
+
ul.css("> li") # the children of ul
|
|
54
|
+
|
|
55
|
+
# XPath 1.0
|
|
56
|
+
doc.xpath("//div[@class='card']//span[@class='price']") # tags
|
|
57
|
+
doc.xpath_values("//a[contains(@href, '/product/')]/@href")
|
|
58
|
+
doc.xpath_value("//dt[.='Price']/following-sibling::dd[1]")
|
|
59
|
+
doc.xpath_value("count(//table//tr)") # "12"
|
|
60
|
+
|
|
61
|
+
# compile once, reuse on any number of pages and threads
|
|
62
|
+
TITLES = fh.XPathQuery("//article//h3/a/@title")
|
|
63
|
+
TITLES.values(doc)
|
|
64
|
+
PRICES = fh.CssQuery("p.price_color")
|
|
65
|
+
PRICES.all(doc)
|
|
66
|
+
|
|
67
|
+
# search by name (any name, case-insensitive) and criteria, as in BeautifulSoup
|
|
68
|
+
pod = doc.find("article", class_="product_pod")
|
|
69
|
+
pod.find("p", class_="price_color").text
|
|
70
|
+
doc.find_all("a", attr=("rel", "next"))
|
|
71
|
+
doc.find("dt", text="Price").find_next_sibling("dd").text
|
|
72
|
+
doc.find("my-widget", id_starts_with="w-")
|
|
73
|
+
# criteria: class_, id, attr=(name, value), text, id_starts_with, text_starts_with, body, body_starts_with, source
|
|
74
|
+
|
|
75
|
+
# reading a tag
|
|
76
|
+
tag.name, tag.depth, tag.index, tag.pos # pos: byte offsets in the UTF-8 source
|
|
77
|
+
tag.text # visible text, entities decoded, whitespace collapsed
|
|
78
|
+
tag.text_line_breaks # <br> and block elements → new lines, <p> and headings → paragraphs
|
|
79
|
+
tag.attr("data-src") # decoded value, "" when missing; attr_raw(): as written
|
|
80
|
+
tag.attrs # {"href": "...", "class": "..."}
|
|
81
|
+
tag.href, tag.src # absolute URLs; href_from(base) / src_from(base)
|
|
82
|
+
tag.mailto, tag.tel # the address of a mailto: link, the number of a tel: link
|
|
83
|
+
tag.body, tag.outer_html, tag.source # inner HTML, the whole element, the opening tag
|
|
84
|
+
|
|
85
|
+
# navigation
|
|
86
|
+
tag.parent, tag.children, tag.next_sibling, tag.previous_sibling, tag.next, tag.previous
|
|
87
|
+
tag.find_parent("table"), tag.find_next("h2"), tag.text_after_until("br")
|
|
88
|
+
|
|
89
|
+
# the whole document
|
|
90
|
+
len(doc), doc[0], doc[-1], list(doc) # every tag, comments included
|
|
91
|
+
```
|
|
92
|
+
|
|
93
|
+
A missing tag is `None`, a missing value is `""`. An invalid selector raises `fh.CssSelectorError`, an invalid XPath
|
|
94
|
+
expression `fh.XPathError` (both are `ValueError`s); `xpath()` on an expression that gives a number or string raises
|
|
95
|
+
`XPathError` too (use `xpath_value`). The string methods compile and cache queries. A `Node` keeps its document alive,
|
|
96
|
+
nodes compare equal when they are the same tag, and are hashable. The GIL is released while a page is parsed, and
|
|
97
|
+
documents and compiled queries can be shared between threads. Type stubs are included.
|
|
98
|
+
|
|
99
|
+
CSS: type, `*`, `#id`, `.class`, attribute selectors with all operators and the `i` / `s` flags, the four combinators,
|
|
100
|
+
selector lists, `:first-child`, `:last-child`, `:only-child`, `:*-of-type`, `:nth-child(An+B [of S])` and its relatives,
|
|
101
|
+
`:not()`, `:is()`, `:where()`, `:has()`, `:empty`, `:root`, `:scope`, `:link`, `:checked`, `:disabled`, `:enabled`,
|
|
102
|
+
`:required`, `:optional`, `:contains("text")`. XPath: all axes except namespace, predicates, `|`, the operators and all
|
|
103
|
+
XPath 1.0 functions, plus `ends-with()`, `lower-case()`, `upper-case()`. HTML rules apply: names are case-insensitive,
|
|
104
|
+
the string-value of an element is its visible text, attribute values and text are decoded. HTML5 optional end tags
|
|
105
|
+
(`<li>`, `<p>`, `<td>`, `<option>` without end tags) give the same elements as in a browser; there is no full HTML5
|
|
106
|
+
tree construction (no implied `<tbody>` or `<html>`), and documents are read-only.
|
|
107
|
+
|
|
108
|
+
## Same results as the .NET version
|
|
109
|
+
|
|
110
|
+
`tests/test_conformance.py` runs the conformance data of the Rust crate through the Python package: on 117 pages
|
|
111
|
+
(fixtures, broken markup, cut-off pages, CR LF, 600-deep nesting, ...) every tag's name, position, text, text with line
|
|
112
|
+
breaks, attributes, navigation and links, and 569 XPath / CSS queries on the document and from three tags of each
|
|
113
|
+
page, must give exactly what z3l.FastHtmlParser (.NET) gives. So the bindings are checked as well as the parser.
|
|
114
|
+
|
|
115
|
+
## Benchmarks
|
|
116
|
+
|
|
117
|
+
`python bench/bench.py` (Python 3.14, AMD Ryzen 7 3800X, Windows; [`bench-results.md`](bench-results.md)). Every
|
|
118
|
+
library must extract the same records before anything is timed.
|
|
119
|
+
|
|
120
|
+
Parse a page:
|
|
121
|
+
|
|
122
|
+
| Page | fasthtmlparser | selectolax (lexbor) | lxml | parsel | BeautifulSoup (lxml) |
|
|
123
|
+
|---|---:|---:|---:|---:|---:|
|
|
124
|
+
| quotes, 11 KB | 19 μs | 95 μs | 495 μs | 499 μs | 4,368 μs |
|
|
125
|
+
| laptops, 38 KB | 53 μs | 306 μs | 1,379 μs | 1,420 μs | 12,192 μs |
|
|
126
|
+
| books, 51 KB | 65 μs | 424 μs | 1,547 μs | 1,616 μs | 15,273 μs |
|
|
127
|
+
| books x10, 510 KB | 652 μs | 3,531 μs | 13,873 μs | 14,886 μs | 15,083 μs\* |
|
|
128
|
+
| books without `</li>` `</p>` | 59 μs | 221 μs | 1,460 μs | 1,593 μs | 14,393 μs |
|
|
129
|
+
|
|
130
|
+
\* BeautifulSoup with lxml stops at the first `</html>`: it finds 20 of the 200 books on that page.
|
|
131
|
+
|
|
132
|
+
Scrape a page: parse and extract the records (books: 20 × title, price, rating; quotes: 10 × text, author, tags).
|
|
133
|
+
|
|
134
|
+
| Task | fasthtmlparser | selectolax | lxml / parsel | BeautifulSoup (lxml) |
|
|
135
|
+
|---|---:|---:|---:|---:|
|
|
136
|
+
| 20 books, CSS | 126 μs | 587 μs | parsel 2,543 μs | 19,006 μs |
|
|
137
|
+
| 20 books, XPath | 217 μs | | lxml 2,715 μs | |
|
|
138
|
+
| 20 books, `find` | 152 μs | | | |
|
|
139
|
+
| 10 quotes | 74 μs | 271 μs | parsel 1,030 μs | 5,984 μs |
|
|
140
|
+
|
|
141
|
+
## Build from source
|
|
142
|
+
|
|
143
|
+
Needs Rust 1.85+ and [maturin](https://www.maturin.rs/):
|
|
144
|
+
|
|
145
|
+
```
|
|
146
|
+
python -m venv .venv && .venv/Scripts/activate # or . .venv/bin/activate
|
|
147
|
+
pip install maturin pytest
|
|
148
|
+
maturin develop --release
|
|
149
|
+
pytest
|
|
150
|
+
```
|
|
151
|
+
|
|
152
|
+
## License
|
|
153
|
+
|
|
154
|
+
Same terms as z3l.FastHtmlParser (see LICENSE.md): free for companies under USD 1M annual revenue and under 100
|
|
155
|
+
people, a commercial license above that. This is source-available, not an OSI open-source license.
|
|
156
|
+
|
|
@@ -0,0 +1,131 @@
|
|
|
1
|
+
# fasthtmlparser
|
|
2
|
+
|
|
3
|
+
[](https://github.com/romanti3172/fasthtmlparser-py/actions/workflows/ci.yml)
|
|
4
|
+
[](https://pypi.org/project/fasthtmlparser/)
|
|
5
|
+
|
|
6
|
+
A fast HTML / XML parser for scraping, with **XPath 1.0** and **CSS selectors**: Python bindings of the
|
|
7
|
+
[Rust port](https://github.com/romanti3172/fasthtmlparser-rs) of
|
|
8
|
+
[z3l.FastHtmlParser](https://github.com/romanti3172/z3l.FastHtmlParser). It is built for extracting data, not for
|
|
9
|
+
building a DOM: a page becomes a flat list of tags that hold only positions in the source; text, attributes and
|
|
10
|
+
entities are read when you ask for them. Parsing is **5–7x faster than selectolax** and **20x faster than lxml**
|
|
11
|
+
(see below).
|
|
12
|
+
|
|
13
|
+
```
|
|
14
|
+
pip install fasthtmlparser
|
|
15
|
+
```
|
|
16
|
+
|
|
17
|
+
Python 3.10 or later. Wheels for Linux (glibc and musl), Windows and macOS, x86-64 and ARM64; no dependencies.
|
|
18
|
+
|
|
19
|
+
```python
|
|
20
|
+
import fasthtmlparser as fh
|
|
21
|
+
|
|
22
|
+
doc = fh.parse(html, url="https://books.toscrape.com/") # str or bytes; url resolves relative links
|
|
23
|
+
|
|
24
|
+
# CSS selectors, as querySelectorAll / querySelector
|
|
25
|
+
for pod in doc.css("article.product_pod"):
|
|
26
|
+
title = pod.css_first("h3 a").attr("title")
|
|
27
|
+
price = pod.css_first("p.price_color").text
|
|
28
|
+
link = pod.css_first("h3 a").href # absolute URL
|
|
29
|
+
ul.css("> li") # the children of ul
|
|
30
|
+
|
|
31
|
+
# XPath 1.0
|
|
32
|
+
doc.xpath("//div[@class='card']//span[@class='price']") # tags
|
|
33
|
+
doc.xpath_values("//a[contains(@href, '/product/')]/@href")
|
|
34
|
+
doc.xpath_value("//dt[.='Price']/following-sibling::dd[1]")
|
|
35
|
+
doc.xpath_value("count(//table//tr)") # "12"
|
|
36
|
+
|
|
37
|
+
# compile once, reuse on any number of pages and threads
|
|
38
|
+
TITLES = fh.XPathQuery("//article//h3/a/@title")
|
|
39
|
+
TITLES.values(doc)
|
|
40
|
+
PRICES = fh.CssQuery("p.price_color")
|
|
41
|
+
PRICES.all(doc)
|
|
42
|
+
|
|
43
|
+
# search by name (any name, case-insensitive) and criteria, as in BeautifulSoup
|
|
44
|
+
pod = doc.find("article", class_="product_pod")
|
|
45
|
+
pod.find("p", class_="price_color").text
|
|
46
|
+
doc.find_all("a", attr=("rel", "next"))
|
|
47
|
+
doc.find("dt", text="Price").find_next_sibling("dd").text
|
|
48
|
+
doc.find("my-widget", id_starts_with="w-")
|
|
49
|
+
# criteria: class_, id, attr=(name, value), text, id_starts_with, text_starts_with, body, body_starts_with, source
|
|
50
|
+
|
|
51
|
+
# reading a tag
|
|
52
|
+
tag.name, tag.depth, tag.index, tag.pos # pos: byte offsets in the UTF-8 source
|
|
53
|
+
tag.text # visible text, entities decoded, whitespace collapsed
|
|
54
|
+
tag.text_line_breaks # <br> and block elements → new lines, <p> and headings → paragraphs
|
|
55
|
+
tag.attr("data-src") # decoded value, "" when missing; attr_raw(): as written
|
|
56
|
+
tag.attrs # {"href": "...", "class": "..."}
|
|
57
|
+
tag.href, tag.src # absolute URLs; href_from(base) / src_from(base)
|
|
58
|
+
tag.mailto, tag.tel # the address of a mailto: link, the number of a tel: link
|
|
59
|
+
tag.body, tag.outer_html, tag.source # inner HTML, the whole element, the opening tag
|
|
60
|
+
|
|
61
|
+
# navigation
|
|
62
|
+
tag.parent, tag.children, tag.next_sibling, tag.previous_sibling, tag.next, tag.previous
|
|
63
|
+
tag.find_parent("table"), tag.find_next("h2"), tag.text_after_until("br")
|
|
64
|
+
|
|
65
|
+
# the whole document
|
|
66
|
+
len(doc), doc[0], doc[-1], list(doc) # every tag, comments included
|
|
67
|
+
```
|
|
68
|
+
|
|
69
|
+
A missing tag is `None`, a missing value is `""`. An invalid selector raises `fh.CssSelectorError`, an invalid XPath
|
|
70
|
+
expression `fh.XPathError` (both are `ValueError`s); `xpath()` on an expression that gives a number or string raises
|
|
71
|
+
`XPathError` too (use `xpath_value`). The string methods compile and cache queries. A `Node` keeps its document alive,
|
|
72
|
+
nodes compare equal when they are the same tag, and are hashable. The GIL is released while a page is parsed, and
|
|
73
|
+
documents and compiled queries can be shared between threads. Type stubs are included.
|
|
74
|
+
|
|
75
|
+
CSS: type, `*`, `#id`, `.class`, attribute selectors with all operators and the `i` / `s` flags, the four combinators,
|
|
76
|
+
selector lists, `:first-child`, `:last-child`, `:only-child`, `:*-of-type`, `:nth-child(An+B [of S])` and its relatives,
|
|
77
|
+
`:not()`, `:is()`, `:where()`, `:has()`, `:empty`, `:root`, `:scope`, `:link`, `:checked`, `:disabled`, `:enabled`,
|
|
78
|
+
`:required`, `:optional`, `:contains("text")`. XPath: all axes except namespace, predicates, `|`, the operators and all
|
|
79
|
+
XPath 1.0 functions, plus `ends-with()`, `lower-case()`, `upper-case()`. HTML rules apply: names are case-insensitive,
|
|
80
|
+
the string-value of an element is its visible text, attribute values and text are decoded. HTML5 optional end tags
|
|
81
|
+
(`<li>`, `<p>`, `<td>`, `<option>` without end tags) give the same elements as in a browser; there is no full HTML5
|
|
82
|
+
tree construction (no implied `<tbody>` or `<html>`), and documents are read-only.
|
|
83
|
+
|
|
84
|
+
## Same results as the .NET version
|
|
85
|
+
|
|
86
|
+
`tests/test_conformance.py` runs the conformance data of the Rust crate through the Python package: on 117 pages
|
|
87
|
+
(fixtures, broken markup, cut-off pages, CR LF, 600-deep nesting, ...) every tag's name, position, text, text with line
|
|
88
|
+
breaks, attributes, navigation and links, and 569 XPath / CSS queries on the document and from three tags of each
|
|
89
|
+
page, must give exactly what z3l.FastHtmlParser (.NET) gives. So the bindings are checked as well as the parser.
|
|
90
|
+
|
|
91
|
+
## Benchmarks
|
|
92
|
+
|
|
93
|
+
`python bench/bench.py` (Python 3.14, AMD Ryzen 7 3800X, Windows; [`bench-results.md`](bench-results.md)). Every
|
|
94
|
+
library must extract the same records before anything is timed.
|
|
95
|
+
|
|
96
|
+
Parse a page:
|
|
97
|
+
|
|
98
|
+
| Page | fasthtmlparser | selectolax (lexbor) | lxml | parsel | BeautifulSoup (lxml) |
|
|
99
|
+
|---|---:|---:|---:|---:|---:|
|
|
100
|
+
| quotes, 11 KB | 19 μs | 95 μs | 495 μs | 499 μs | 4,368 μs |
|
|
101
|
+
| laptops, 38 KB | 53 μs | 306 μs | 1,379 μs | 1,420 μs | 12,192 μs |
|
|
102
|
+
| books, 51 KB | 65 μs | 424 μs | 1,547 μs | 1,616 μs | 15,273 μs |
|
|
103
|
+
| books x10, 510 KB | 652 μs | 3,531 μs | 13,873 μs | 14,886 μs | 15,083 μs\* |
|
|
104
|
+
| books without `</li>` `</p>` | 59 μs | 221 μs | 1,460 μs | 1,593 μs | 14,393 μs |
|
|
105
|
+
|
|
106
|
+
\* BeautifulSoup with lxml stops at the first `</html>`: it finds 20 of the 200 books on that page.
|
|
107
|
+
|
|
108
|
+
Scrape a page: parse and extract the records (books: 20 × title, price, rating; quotes: 10 × text, author, tags).
|
|
109
|
+
|
|
110
|
+
| Task | fasthtmlparser | selectolax | lxml / parsel | BeautifulSoup (lxml) |
|
|
111
|
+
|---|---:|---:|---:|---:|
|
|
112
|
+
| 20 books, CSS | 126 μs | 587 μs | parsel 2,543 μs | 19,006 μs |
|
|
113
|
+
| 20 books, XPath | 217 μs | | lxml 2,715 μs | |
|
|
114
|
+
| 20 books, `find` | 152 μs | | | |
|
|
115
|
+
| 10 quotes | 74 μs | 271 μs | parsel 1,030 μs | 5,984 μs |
|
|
116
|
+
|
|
117
|
+
## Build from source
|
|
118
|
+
|
|
119
|
+
Needs Rust 1.85+ and [maturin](https://www.maturin.rs/):
|
|
120
|
+
|
|
121
|
+
```
|
|
122
|
+
python -m venv .venv && .venv/Scripts/activate # or . .venv/bin/activate
|
|
123
|
+
pip install maturin pytest
|
|
124
|
+
maturin develop --release
|
|
125
|
+
pytest
|
|
126
|
+
```
|
|
127
|
+
|
|
128
|
+
## License
|
|
129
|
+
|
|
130
|
+
Same terms as z3l.FastHtmlParser (see LICENSE.md): free for companies under USD 1M annual revenue and under 100
|
|
131
|
+
people, a commercial license above that. This is source-available, not an OSI open-source license.
|