Spaces:
Runtime error
Runtime error
File size: 5,402 Bytes
f5ec40e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 |
import math
import os
import bs4
import pandas as pd
from bs4 import BeautifulSoup
def parse_section(nodes: list[bs4.element.NavigableString]) -> str:
section = []
for node in nodes:
if node.name == "table":
node_text = pd.read_html(node.prettify())[0].to_markdown(index=False, tablefmt="github")
elif node.name == "script":
continue
else:
node_text = node.text
section.append(node_text)
section = "".join(section)
return section
class Parser:
def __init__(
self,
soup: BeautifulSoup,
base_url: str,
filename: str,
min_section_length: int = 100,
max_section_length: int = 2000,
):
self.soup = soup
self.base_url = base_url
self.filename = filename
self.min_section_length = min_section_length
self.max_section_length = max_section_length
def parse(self) -> tuple[list[str], list[str], list[str]]:
...
def find_sections(self) -> bs4.element.ResultSet:
...
def build_url(self, suffix: str) -> str:
...
class SphinxParser(Parser):
def parse(self) -> tuple[list[str], list[str], list[str]]:
found = self.find_sections()
sections = []
urls = []
names = []
for i in range(len(found)):
section_found = found[i]
section_soup = section_found.parent.parent
section_href = section_soup.find_all("a", href=True, class_="headerlink")
# If sections has subsections, keep only the part before the first subsection
if len(section_href) > 1 and section_soup.section is not None:
section_siblings = list(section_soup.section.previous_siblings)[::-1]
section = parse_section(section_siblings)
else:
section = parse_section(section_soup.children)
# Remove special characters, plus newlines in some url and section names.
section = section.strip()
url = section_found["href"].strip().replace("\n", "")
name = section_found.parent.text.strip()[:-1].replace("\n", "")
url = self.build_url(url)
# If text is too long, split into chunks of equal sizes
if len(section) > self.max_section_length:
n_chunks = math.ceil(len(section) / float(self.max_section_length))
separator_index = math.floor(len(section) / n_chunks)
section_chunks = [section[separator_index * i : separator_index * (i + 1)] for i in range(n_chunks)]
url_chunks = [url] * n_chunks
name_chunks = [name] * n_chunks
sections.extend(section_chunks)
urls.extend(url_chunks)
names.extend(name_chunks)
# If text is not too short, add in 1 chunk
elif len(section) > self.min_section_length:
sections.append(section)
urls.append(url)
names.append(name)
return sections, urls, names
def find_sections(self) -> bs4.element.ResultSet:
return self.soup.find_all("a", href=True, class_="headerlink")
def build_url(self, suffix: str) -> str:
return self.base_url + self.filename + suffix
class HuggingfaceParser(Parser):
def parse(self) -> tuple[list[str], list[str], list[str]]:
found = self.find_sections()
sections = []
urls = []
names = []
for i in range(len(found)):
section_href = found[i].find("a", href=True, class_="header-link")
section_nodes = []
for element in found[i].find_next_siblings():
if i + 1 < len(found) and element == found[i + 1]:
break
section_nodes.append(element)
section = parse_section(section_nodes)
# Remove special characters, plus newlines in some url and section names.
section = section.strip()
url = section_href["href"].strip().replace("\n", "")
name = found[i].text.strip().replace("\n", "")
url = self.build_url(url)
# If text is too long, split into chunks of equal sizes
if len(section) > self.max_section_length:
n_chunks = math.ceil(len(section) / float(self.max_section_length))
separator_index = math.floor(len(section) / n_chunks)
section_chunks = [section[separator_index * i : separator_index * (i + 1)] for i in range(n_chunks)]
url_chunks = [url] * n_chunks
name_chunks = [name] * n_chunks
sections.extend(section_chunks)
urls.extend(url_chunks)
names.extend(name_chunks)
# If text is not too short, add in 1 chunk
elif len(section) > self.min_section_length:
sections.append(section)
urls.append(url)
names.append(name)
return sections, urls, names
def find_sections(self) -> bs4.element.ResultSet:
return self.soup.find_all(["h1", "h2", "h3"], class_="relative group")
def build_url(self, suffix: str) -> str:
# The splitext is to remove the .html extension
return self.base_url + os.path.splitext(self.filename)[0] + suffix
|