dribdat/dribdat/apifetch.py

# -*- coding: utf-8 -*-
"""Collecting data from third party API repositories."""

import re
import requests
import bleach
import logging
from pyquery import PyQuery as pq  # noqa: N813
from base64 import b64decode
from flask_misaka import markdown
from bleach.sanitizer import ALLOWED_ATTRIBUTES
from urllib.parse import quote_plus
from .apievents import (
    fetch_commits_github,
    fetch_commits_gitlab,
    fetch_commits_gitea,
)
from .utils import (
    sanitize_url, load_presets, load_yaml_presets, fix_relative_links
)
from future.standard_library import install_aliases
install_aliases()

# In seconds, how long to wait for API response
REQUEST_TIMEOUT = 10


def FetchStageConfig(url, top_element='stages', by_col='name'):
    """Download a remote YAML stages configuration."""
    if not url.startswith('http:') and not url.startswith('https:'):
        logging.info("Loading stages from file")
        return load_yaml_presets(top_element, by_col, url)
    logging.info("Loading stages from URL")
    data = requests.get(url, timeout=REQUEST_TIMEOUT)
    if data.text.find('stages:') < 0:
        logging.debug("No stage data", data.text)
        return {}
    blob = data.text
    return load_presets(blob, top_element, by_col)


def FetchGiteaProject(project_url):
    """Download data from Codeberg, a large Gitea site."""
    # Docs: https://codeberg.org/api/swagger
    site_root = "https://codeberg.org"
    url_q = quote_plus(project_url, '/')
    api_repos = site_root + "/api/v1/repos/%s" % url_q
    api_content = api_repos + "/contents"
    # Collect basic data
    logging.info("Fetching Gitea", url_q)
    data = requests.get(api_repos, timeout=REQUEST_TIMEOUT)
    if data.text.find('{') < 0:
        logging.debug("No data", data.text)
        return {}
    json = data.json()
    if 'name' not in json:
        logging.debug("Invalid data", data.text)
        return {}
    # Collect the README
    data = requests.get(api_content, timeout=REQUEST_TIMEOUT)
    readme = ""
    if not data.text.find('{') < 0:
        readmeurl = None
        for repo_file in data.json():
            if 'readme' in repo_file['name'].lower():
                readmeurl = repo_file['download_url']
                readmedata = requests.get(readmeurl, timeout=REQUEST_TIMEOUT)
                readme = readmedata.text
                break
        if readmeurl is None:
            logging.info("Could not find README", url_q)
    issuesurl = ''
    if json['has_issues']:
        issuesurl = json['html_url'] + '/issues'
    return {
        'type': 'Gitea',
        'name': json['name'],
        'summary': json['description'],
        'description': readme,
        'source_url': json['html_url'],
        'image_url': json['avatar_url'] or json['owner']['avatar_url'],
        'contact_url': issuesurl,
        'commits': fetch_commits_gitea(url_q)
    }


def FetchGitlabProject(project_url):
    """Download data from GitLab."""
    WEB_BASE = "https://gitlab.com"
    API_BASE = WEB_BASE + "/api/v4/projects/%s"
    url_q = quote_plus(project_url)
    # Collect basic data
    logging.info("Fetching GitLab", url_q)
    data = requests.get(API_BASE % url_q, timeout=REQUEST_TIMEOUT)
    if data.text.find('{') < 0:
        logging.debug("No data", data.text)
        return {}
    json = data.json()
    if 'name' not in json:
        logging.debug("Invalid data", data.text)
        return {}
    # Collect the README
    readmeurl = json['readme_url'] + '?inline=false'
    readmeurl = readmeurl.replace('-/blob/', '-/raw/')
    readmedata = requests.get(readmeurl, timeout=REQUEST_TIMEOUT)
    readme = readmedata.text or ""
    return {
        'type': 'GitLab',
        'name': json['name'],
        'summary': json['description'],
        'description': readme,
        'source_url': json['web_url'],
        'image_url': json['avatar_url'],
        'contact_url': json['web_url'] + '/issues',
        'commits': fetch_commits_gitlab(json['id'])
    }


def FetchGitlabAvatar(email):
    """Download a user avatar from GitLab."""
    apiurl = "https://gitlab.com/api/v4/avatar?email=%s&size=80"
    data = requests.get(apiurl % email, timeout=REQUEST_TIMEOUT)
    if data.text.find('{') < 0:
        logging.debug("No data", data.text)
        return None
    json = data.json()
    if 'avatar_url' not in json:
        return None
    return json['avatar_url']


def FetchGithubProject(project_url):
    """Download data from GitHub."""
    API_BASE = "https://api.github.com/repos/%s"
    logging.info("Fetching GitHub", project_url)
    data = requests.get(API_BASE % project_url, timeout=REQUEST_TIMEOUT)
    if data.text.find('{') < 0:
        logging.debug("No data", data.text)
        return {}
    json = data.json()
    if 'name' not in json or 'full_name' not in json:
        logging.debug("Invalid data", data.text)
        return {}
    repo_full_name = json['full_name']
    default_branch = json['default_branch'] or 'main'
    readmeurl = "%s/readme" % (API_BASE % project_url)
    readmedata = requests.get(readmeurl, timeout=REQUEST_TIMEOUT)
    readme = ''
    if readmedata.text.find('{') < 0:
        logging.debug("No readme", data.text)
    else:
        readme = readmedata.json()
    if 'content' not in readme:
        readme = ''
    else:
        readme = b64decode(readme['content']).decode('utf-8')
        # Fix relative links in text
        imgroot = "https://raw.githubusercontent.com"
        readme = fix_relative_links(readme, imgroot, repo_full_name, default_branch)
    return {
        'type': 'GitHub',
        'name': json['name'],
        'summary': json['description'],
        'description': readme,
        'webpage_url': json['homepage'],
        'source_url': json['html_url'],
        'image_url': json['owner']['avatar_url'],
        'contact_url': json['html_url'] + '/issues',
        'download_url': json['html_url'] + '/releases',
        'commits': fetch_commits_github(repo_full_name)
    }


def FetchBitbucketProject(project_url):
    """Download data from Bitbucket."""
    WEB_BASE = "https://bitbucket.org/%s"
    API_BASE = "https://api.bitbucket.org/2.0/repositories/%s"
    logging.info("Fetching Bitbucket", project_url)
    data = requests.get(API_BASE % project_url, timeout=REQUEST_TIMEOUT)
    if data.text.find('{') < 0:
        logging.debug('No data at', project_url)
        return {}
    json = data.json()
    if 'name' not in json:
        logging.debug('Invalid format at', project_url)
        return {}
    readme = ''
    for docext in ['.md', '.rst', '.txt', '']:
        readmedata = requests.get(
            API_BASE % project_url + '/src/HEAD/README.md',
            timeout=REQUEST_TIMEOUT)
        if readmedata.text.find('{"type":"error"') != 0:
            readme = readmedata.text
            break
    web_url = WEB_BASE % project_url
    contact_url = json['website'] or web_url
    if json['has_issues']:
        contact_url = "%s/issues" % web_url
    image_url = ''
    if 'project' in json and \
            'links' in json['project'] \
            and 'avatar' in json['project']['links']:
        image_url = json['project']['links']['avatar']['href']
    elif 'links' in json and 'avatar' in json['links']:
        image_url = json['links']['avatar']['href']
    return {
        'type': 'Bitbucket',
        'name': json['name'],
        'summary': json['description'],
        'description': readme,
        'webpage_url': json['website'],
        'source_url': web_url,
        'image_url': image_url,
        'contact_url': contact_url,
    }


def FetchDataProject(datapackage_url):
    """Try to load a Data Package formatted JSON file."""
    # TODO: use frictionlessdata library!
    project_url = datapackage_url.replace('datapackage.json', '')
    project_url = sanitize_url(project_url) + 'datapackage.json'
    data = requests.get(project_url, timeout=REQUEST_TIMEOUT)
    # TODO: treat dribdat events as special
    logging.info("Fetching Data Package", project_url)
    if data.text.find('{') < 0:
        logging.debug('No data at', project_url)
        return {}
    json = data.json()
    if 'name' not in json or 'title' not in json:
        logging.debug('Invalid format at', project_url)
        return {}
    try:
        text_content = parse_data_package(json)
    except KeyError:
        text_content = '(Could not parse Data Package contents)'
    if 'homepage' in json:
        contact_url = json['homepage'] or ''
    elif 'maintainers' in json and \
            len(json['maintainers']) > 0 and \
            'web' in json['maintainers'][0]:
        contact_url = json['maintainers'][0]['web']
    return {
        'type': 'Data Package',
        'name': json['name'],
        'summary': json['title'],
        'description': text_content,
        'source_url': project_url,
        'logo_icon': 'box-open',
        'contact_url': contact_url,
    }


def parse_data_package(json):
    """Extract contents of a Data Package."""
    text_content = ''
    if 'description' in json:
        text_content = json['description'] + '\n\n'
    if 'resources' in json:
        text_content = text_content + '\n### Resources\n\n'
        for r in json['resources']:
            rn = r['name']
            if 'path' in r:
                rn = "[%s](%s)" % (rn, r['path'])
            text_content = text_content + '- ' + rn + '\n'
    if 'sources' in json:
        text_content = text_content + '\n### Sources\n\n'
        for r in json['sources']:
            rn = r['title']
            if 'path' in r:
                rn = "[%s](%s)" % (rn, r['path'])
            text_content = text_content + '- ' + rn + '\n'
    if text_content == '':
        raise KeyError('No content')
    return text_content


def FetchDribdatProject(dribdat_url):
    """Try to load a Dribdat project from a remote page."""
    project_url = dribdat_url.replace('/project/', '/api/project/') 
    project_url = sanitize_url(project_url) + '?full=1'
    data = requests.get(project_url, timeout=REQUEST_TIMEOUT)
    # TODO: treat dribdat events as special
    logging.info("Fetching Dribdat site", project_url)
    if data.text.find('{') < 0:
        logging.debug('No data at', project_url)
        return {}
    json = data.json()
    if 'project' not in json or 'event' not in json:
        logging.debug('Invalid format at', project_url)
        return {}
    projectdata = json['project']
    projectdata['type'] = 'Dribdat'
    projectdata['description'] = projectdata['longtext']
    return projectdata



# Basis: https://github.com/mozilla/bleach/blob/master/bleach/sanitizer.py#L16
ALLOWED_HTML_TAGS = [
    'acronym', 'a', 'blockquote', 'li', 'abbr', 
    'strong', 'b', 'i', 'ul', 'ol', 'code', 'em',
    'img', 'font', 'center', 'sub', 'sup', 'pre',
    'table', 'tr', 'thead', 'tbody', 'td',
    'h1', 'h2', 'h3', 'h4', 'h5',
    'p', 'u'
]
ALLOWED_HTML_ATTR = ALLOWED_ATTRIBUTES
ALLOWED_HTML_ATTR['h1'] = ['id']
ALLOWED_HTML_ATTR['h2'] = ['id']
ALLOWED_HTML_ATTR['h3'] = ['id']
ALLOWED_HTML_ATTR['h4'] = ['id']
ALLOWED_HTML_ATTR['h5'] = ['id']
ALLOWED_HTML_ATTR['a'] = ['href', 'title', 'class', 'name']
ALLOWED_HTML_ATTR['img'] = ['src', 'width', 'height', 'alt', 'class']
ALLOWED_HTML_ATTR['font'] = ['color']


def RequestRemoteContent(project_url):
    try:
        # TODO: the admin should be able to whitelist a range of allowed
        # online resources controlling the domains from which we can 
        # fetch remote content.
        project_url = sanitize_url(project_url)
        logging.info("Fetching", project_url)
        data = requests.get(project_url, timeout=REQUEST_TIMEOUT)
        return data.text or None
    except requests.exceptions.RequestException:
        logging.warning("Could not connect to %s" % project_url)
        return None


def FetchWebProject(project_url):
    """Parse a remote Document, wiki or website URL."""
    
    datatext = RequestRemoteContent(project_url)
    if datatext is None: return {}

    # Google Document
    if project_url.startswith('https://docs.google.com/document'):
        return FetchWebGoogleDoc(datatext, project_url)
    # Instructables
    elif project_url.startswith('https://www.instructables.com/'):
        return FetchWebInstructables(datatext, project_url)
    # CodiMD / HackMD
    elif datatext.find('<div id="doc" ') > 0:
        return FetchWebCodiMD(datatext, project_url)
    # DokuWiki
    elif datatext.find('<meta name="generator" content="DokuWiki"/>') > 0:
        return FetchWebDokuWiki(datatext, project_url)
    # Etherpad
    elif datatext.find('pad.importExport.exportetherpad') > 0:
        return FetchWebEtherpad(datatext, project_url)


def FetchWebGoogleDoc(text, url):
    """Help extract data from a Google doc."""
    doc = pq(text)
    doc("style").remove()
    ptitle = doc("div#title") or doc("div#header")
    if len(ptitle) < 1:
        return {}
    content = doc("div#contents")
    if len(content) < 1:
        return {}
    content = content.html().strip()
    if not content or len(content) < 1:
        return {}
    html_content = bleach.clean(content, strip=True,
                                tags=frozenset(ALLOWED_HTML_TAGS),
                                attributes=ALLOWED_HTML_ATTR)
    obj = {}
    # {
    #     'type': 'Google', ...
    #     'name': name,
    #     'summary': summary,
    #     'description': html_content,
    #     'image_url': image_url
    #     'source_url': project_url,
    # }
    obj['type'] = 'Google Docs'
    obj['name'] = ptitle.text()
    obj['description'] = html_content
    obj['source_url'] = url
    obj['logo_icon'] = 'paperclip'
    return obj


def FetchWebCodiMD(text, url):
    """Help extract data from CodiMD."""
    doc = pq(text)
    ptitle = doc("title")
    if len(ptitle) < 1:
        return {}
    content = doc("div#doc").html()
    if len(content) < 1:
        return {}
    obj = {}
    obj['type'] = 'Markdown'
    obj['name'] = ptitle.text()
    obj['description'] = markdown(content)
    obj['source_url'] = url
    obj['logo_icon'] = 'outdent'
    return obj


def FetchWebDokuWiki(text, url):
    """Help extract data from DokuWiki."""
    doc = pq(text)
    ptitle = doc("span.pageId")
    if len(ptitle) < 1:
        return {}
    content = doc("div.dw-content")
    if len(content) < 1:
        return {}
    html_content = bleach.clean(content.html().strip(), strip=True,
                                tags=ALLOWED_HTML_TAGS,
                                attributes=ALLOWED_HTML_ATTR)
    obj = {}
    obj['type'] = 'DokuWiki'
    obj['name'] = ptitle.text().replace('project:', '')
    obj['description'] = html_content
    obj['source_url'] = url
    obj['logo_icon'] = 'list-ul'
    return obj


def FetchWebEtherpad(text, url):
    """Help extract data from Etherpad Lite."""
    ptitle = url.split('/')[-1]
    if len(ptitle) < 1:
        return {}
    text_content = requests.get(
        "%s/export/txt" % url,
        timeout=REQUEST_TIMEOUT).text
    obj = {}
    obj['type'] = 'Etherpad'
    obj['name'] = ptitle.replace('_', ' ')
    obj['description'] = text_content
    obj['source_url'] = url
    obj['logo_icon'] = 'pen'
    return obj


def FetchWebInstructables(text, url):
    """Help extract data from Instructables."""
    doc = pq(text)
    ptitle = doc(".header-title")
    content = doc(".main-content")
    if len(content) < 1 or len(ptitle) < 1:
        return {}
    html_content = ParseInstructablesPage(content)
    obj = {}
    obj['type'] = 'Instructables'
    obj['name'] = ptitle.text()
    obj['description'] = html_content
    obj['source_url'] = url
    obj['logo_icon'] = 'wrench'
    return obj


def FetchWebGitHub(url):
    """Grab a Markdown source from a GitHub link."""
    if not url.endswith('.md') or not '/blob/' in url:
        return {}
    filename = url.split('/')[-1].replace('.md', '')
    rawurl = url.replace('/blob/', '/raw/').replace("https://github.com/", '')
    rawdata = requests.get("https://github.com/" + rawurl, timeout=REQUEST_TIMEOUT)
    text_content = rawdata.text or ""
    return {
        'type': 'Markdown',
        'name': filename,
        'description': text_content,
        'source_url': url,
        'logo_icon': 'outdent',
    }


def FetchWebGitHubGist(url):
    """Grab a Markdown source from a GitHub Gist link."""
    rawurl = url.replace("https://gist.github.com/", '') + '/raw'
    rawdata = requests.get("https://gist.githubusercontent.com/" + rawurl, timeout=REQUEST_TIMEOUT)
    text_content = rawdata.text or ""
    return {
        'type': 'Markdown',
        'name': 'Gist',
        'description': text_content,
        'source_url': url,
        'logo_icon': 'outdent',
    }


def ParseInstructablesPage(content):
    """Create an HTML summary of content."""
    html_content = ""
    for step in content.find(".step"):
        step_title = pq(step).find('.step-title')
        if step_title is not None:
            html_content += '<h3>' + step_title.text() + '</h3>'
        # Grab photos
        for img in pq(step).find('noscript'):
            if '{{ file' not in pq(img).html():
                html_content += pq(img).html()
        # Iterate through body
        step_content = pq(step).find('.step-body')
        if step_content is None:
            continue
        for elem in pq(step_content).children():
            elem_tag, p = ParseInstructablesElement(elem)
            if elem_tag is None:
                continue
            html_content += '<%s>%s</%s>' % (elem_tag, p, elem_tag)
    return html_content


def ParseInstructablesElement(elem):
    """Check and return minimal contents."""
    if elem.tag == 'pre':
        if elem.text is None:
            return None, None
        return 'pre', elem.text
    else:
        p = pq(elem).html()
        if p is None:
            return None, None
        p = bleach.clean(p.strip(), strip=True,
                         tags=ALLOWED_HTML_TAGS,
                         attributes=ALLOWED_HTML_ATTR)
        return elem.tag, p