JobSpy/src/jobspy/scrapers/indeed/__init__.py

436 lines
15 KiB
Python

"""
jobspy.scrapers.indeed
~~~~~~~~~~~~~~~~~~~
This module contains routines to scrape Indeed.
"""
from __future__ import annotations
import math
from typing import Tuple
from datetime import datetime
from concurrent.futures import ThreadPoolExecutor, Future
import requests
from .. import Scraper, ScraperInput, Site
from ..utils import (
extract_emails_from_text,
get_enum_from_job_type,
markdown_converter,
logger,
)
from ...jobs import (
JobPost,
Compensation,
CompensationInterval,
Location,
JobResponse,
JobType,
DescriptionFormat,
)
class IndeedScraper(Scraper):
def __init__(self, proxy: str | None = None):
"""
Initializes IndeedScraper with the Indeed API url
"""
self.scraper_input = None
self.jobs_per_page = 100
self.num_workers = 10
self.seen_urls = set()
self.headers = None
self.api_country_code = None
self.base_url = None
self.api_url = "https://apis.indeed.com/graphql"
site = Site(Site.INDEED)
super().__init__(site, proxy=proxy)
def scrape(self, scraper_input: ScraperInput) -> JobResponse:
"""
Scrapes Indeed for jobs with scraper_input criteria
:param scraper_input:
:return: job_response
"""
self.scraper_input = scraper_input
domain, self.api_country_code = self.scraper_input.country.indeed_domain_value
self.base_url = f"https://{domain}.indeed.com"
self.headers = self.api_headers.copy()
self.headers["indeed-co"] = self.scraper_input.country.indeed_domain_value
job_list = []
page = 1
cursor = None
offset_pages = math.ceil(self.scraper_input.offset / 100)
for _ in range(offset_pages):
logger.info(f"Indeed skipping search page: {page}")
__, cursor = self._scrape_page(cursor)
if not __:
logger.info(f"Indeed found no jobs on page: {page}")
break
while len(self.seen_urls) < scraper_input.results_wanted:
logger.info(f"Indeed search page: {page}")
jobs, cursor = self._scrape_page(cursor)
if not jobs:
logger.info(f"Indeed found no jobs on page: {page}")
break
job_list += jobs
page += 1
return JobResponse(jobs=job_list[: scraper_input.results_wanted])
def _scrape_page(self, cursor: str | None) -> Tuple[list[JobPost], str | None]:
"""
Scrapes a page of Indeed for jobs with scraper_input criteria
:param cursor:
:return: jobs found on page, next page cursor
"""
jobs = []
new_cursor = None
filters = self._build_filters()
search_term = self.scraper_input.search_term.replace('"', '\\"') if self.scraper_input.search_term else ""
query = self.job_search_query.format(
what=(
f'what: "{search_term}"'
if search_term
else ""
),
location=(
f'location: {{where: "{self.scraper_input.location}", radius: {self.scraper_input.distance}, radiusUnit: MILES}}'
if self.scraper_input.location
else ""
),
dateOnIndeed=self.scraper_input.hours_old,
cursor=f'cursor: "{cursor}"' if cursor else "",
filters=filters,
)
payload = {
"query": query,
}
api_headers = self.api_headers.copy()
api_headers["indeed-co"] = self.api_country_code
response = requests.post(
self.api_url,
headers=api_headers,
json=payload,
proxies=self.proxy,
timeout=10,
)
if response.status_code != 200:
logger.info(
f"Indeed responded with status code: {response.status_code} (submit GitHub issue if this appears to be a bug)"
)
return jobs, new_cursor
data = response.json()
jobs = data["data"]["jobSearch"]["results"]
new_cursor = data["data"]["jobSearch"]["pageInfo"]["nextCursor"]
with ThreadPoolExecutor(max_workers=self.num_workers) as executor:
job_results: list[Future] = [
executor.submit(self._process_job, job["job"]) for job in jobs
]
job_list = [result.result() for result in job_results if result.result()]
return job_list, new_cursor
def _build_filters(self):
"""
Builds the filters dict for job type/is_remote. If hours_old is provided, composite filter for job_type/is_remote is not possible.
IndeedApply: filters: { keyword: { field: "indeedApplyScope", keys: ["DESKTOP"] } }
"""
filters_str = ""
if self.scraper_input.hours_old:
filters_str = """
filters: {{
date: {{
field: "dateOnIndeed",
start: "{start}h"
}}
}}
""".format(
start=self.scraper_input.hours_old
)
if self.scraper_input.easy_apply:
filters_str = """
filters: {
keyword: {
field: "indeedApplyScope",
keys: ["DESKTOP"]
}
}
"""
if self.scraper_input.job_type or self.scraper_input.is_remote:
job_type_key_mapping = {
JobType.FULL_TIME: "CF3CP",
JobType.PART_TIME: "75GKK",
JobType.CONTRACT: "NJXCK",
JobType.INTERNSHIP: "VDTG7",
}
keys = []
if self.scraper_input.job_type:
key = job_type_key_mapping[self.scraper_input.job_type]
keys.append(key)
if self.scraper_input.is_remote:
keys.append("DSQF7")
if keys:
keys_str = '", "'.join(keys) # Prepare your keys string
filters_str = f"""
filters: {{
composite: {{
filters: [{{
keyword: {{
field: "attributes",
keys: ["{keys_str}"]
}}
}}]
}}
}}
"""
return filters_str
def _process_job(self, job: dict) -> JobPost | None:
"""
Parses the job dict into JobPost model
:param job: dict to parse
:return: JobPost if it's a new job
"""
job_url = f'{self.base_url}/viewjob?jk={job["key"]}'
if job_url in self.seen_urls:
return
self.seen_urls.add(job_url)
description = job["description"]["html"]
if self.scraper_input.description_format == DescriptionFormat.MARKDOWN:
description = markdown_converter(description)
job_type = self._get_job_type(job["attributes"])
timestamp_seconds = job["datePublished"] / 1000
date_posted = datetime.fromtimestamp(timestamp_seconds).strftime("%Y-%m-%d")
employer = job["employer"].get("dossier") if job["employer"] else None
employer_details = employer.get("employerDetails", {}) if employer else {}
rel_url = job["employer"]["relativeCompanyPageUrl"] if job["employer"] else None
return JobPost(
id=str(job["key"]),
title=job["title"],
description=description,
company_name=job["employer"].get("name") if job.get("employer") else None,
company_url=(f"{self.base_url}{rel_url}" if job["employer"] else None),
company_url_direct=(
employer["links"]["corporateWebsite"] if employer else None
),
location=Location(
city=job.get("location", {}).get("city"),
state=job.get("location", {}).get("admin1Code"),
country=job.get("location", {}).get("countryCode"),
),
job_type=job_type,
compensation=self._get_compensation(job),
date_posted=date_posted,
job_url=job_url,
job_url_direct=(
job["recruit"].get("viewJobUrl") if job.get("recruit") else None
),
emails=extract_emails_from_text(description) if description else None,
is_remote=self._is_job_remote(job, description),
company_addresses=(
employer_details["addresses"][0]
if employer_details.get("addresses")
else None
),
company_industry=(
employer_details["industry"]
.replace("Iv1", "")
.replace("_", " ")
.title()
if employer_details.get("industry")
else None
),
company_num_employees=employer_details.get("employeesLocalizedLabel"),
company_revenue=employer_details.get("revenueLocalizedLabel"),
company_description=employer_details.get("briefDescription"),
ceo_name=employer_details.get("ceoName"),
ceo_photo_url=employer_details.get("ceoPhotoUrl"),
logo_photo_url=(
employer["images"].get("squareLogoUrl")
if employer and employer.get("images")
else None
),
banner_photo_url=(
employer["images"].get("headerImageUrl")
if employer and employer.get("images")
else None
),
)
@staticmethod
def _get_job_type(attributes: list) -> list[JobType]:
"""
Parses the attributes to get list of job types
:param attributes:
:return: list of JobType
"""
job_types: list[JobType] = []
for attribute in attributes:
job_type_str = attribute["label"].replace("-", "").replace(" ", "").lower()
job_type = get_enum_from_job_type(job_type_str)
if job_type:
job_types.append(job_type)
return job_types
@staticmethod
def _get_compensation(job: dict) -> Compensation | None:
"""
Parses the job to get compensation
:param job:
:param job:
:return: compensation object
"""
comp = job["compensation"]["baseSalary"]
if not comp:
return None
interval = IndeedScraper._get_compensation_interval(comp["unitOfWork"])
if not interval:
return None
min_range = comp["range"].get("min")
max_range = comp["range"].get("max")
return Compensation(
interval=interval,
min_amount=round(min_range, 2) if min_range is not None else None,
max_amount=round(max_range, 2) if max_range is not None else None,
currency=job["compensation"]["currencyCode"],
)
@staticmethod
def _is_job_remote(job: dict, description: str) -> bool:
"""
Searches the description, location, and attributes to check if job is remote
"""
remote_keywords = ["remote", "work from home", "wfh"]
is_remote_in_attributes = any(
any(keyword in attr["label"].lower() for keyword in remote_keywords)
for attr in job["attributes"]
)
is_remote_in_description = any(
keyword in description.lower() for keyword in remote_keywords
)
is_remote_in_location = any(
keyword in job["location"]["formatted"]["long"].lower()
for keyword in remote_keywords
)
return (
is_remote_in_attributes or is_remote_in_description or is_remote_in_location
)
@staticmethod
def _get_compensation_interval(interval: str) -> CompensationInterval:
interval_mapping = {
"DAY": "DAILY",
"YEAR": "YEARLY",
"HOUR": "HOURLY",
"WEEK": "WEEKLY",
"MONTH": "MONTHLY",
}
mapped_interval = interval_mapping.get(interval.upper(), None)
if mapped_interval and mapped_interval in CompensationInterval.__members__:
return CompensationInterval[mapped_interval]
else:
raise ValueError(f"Unsupported interval: {interval}")
api_headers = {
"Host": "apis.indeed.com",
"content-type": "application/json",
"indeed-api-key": "161092c2017b5bbab13edb12461a62d5a833871e7cad6d9d475304573de67ac8",
"accept": "application/json",
"indeed-locale": "en-US",
"accept-language": "en-US,en;q=0.9",
"user-agent": "Mozilla/5.0 (iPhone; CPU iPhone OS 16_6_1 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E148 Indeed App 193.1",
"indeed-app-info": "appv=193.1; appid=com.indeed.jobsearch; osv=16.6.1; os=ios; dtype=phone",
}
job_search_query = """
query GetJobData {{
jobSearch(
{what}
{location}
includeSponsoredResults: NONE
limit: 100
sort: DATE
{cursor}
{filters}
) {{
pageInfo {{
nextCursor
}}
results {{
trackingKey
job {{
key
title
datePublished
dateOnIndeed
description {{
html
}}
location {{
countryName
countryCode
admin1Code
city
postalCode
streetAddress
formatted {{
short
long
}}
}}
compensation {{
baseSalary {{
unitOfWork
range {{
... on Range {{
min
max
}}
}}
}}
currencyCode
}}
attributes {{
key
label
}}
employer {{
relativeCompanyPageUrl
name
dossier {{
employerDetails {{
addresses
industry
employeesLocalizedLabel
revenueLocalizedLabel
briefDescription
ceoName
ceoPhotoUrl
}}
images {{
headerImageUrl
squareLogoUrl
}}
links {{
corporateWebsite
}}
}}
}}
recruit {{
viewJobUrl
detailedSalary
workSchedule
}}
}}
}}
}}
}}
"""