Live scraping updates in pipeline UI (#100)

* initial commit

* fix clear script

* cancelling pipelines

* formatting
This commit is contained in:
Shaheer Sarfaraz
2026-02-07 22:44:00 +00:00
committed by GitHub
parent 60788b0f6a
commit a409aa5ee0
28 changed files with 1425 additions and 84 deletions
+27
View File
@@ -1,9 +1,12 @@
import csv
import json
import os
from pathlib import Path
from jobspy import scrape_jobs
PROGRESS_PREFIX = "JOBOPS_PROGRESS "
def _env_str(name: str, default: str) -> str:
value = os.getenv(name)
@@ -27,6 +30,11 @@ def _env_bool(name: str, default: bool) -> bool:
return value.strip().lower() in ("1", "true", "yes", "y", "on")
def _emit_progress(event: str, payload: dict) -> None:
serialized = json.dumps({"event": event, **payload}, ensure_ascii=True)
print(f"{PROGRESS_PREFIX}{serialized}", flush=True)
def _parse_sites(raw: str) -> list[str]:
return [s.strip() for s in raw.split(",") if s.strip()]
@@ -40,6 +48,8 @@ def main() -> int:
country_indeed = _env_str("JOBSPY_COUNTRY_INDEED", "UK")
linkedin_fetch_description = _env_bool("JOBSPY_LINKEDIN_FETCH_DESCRIPTION", True)
is_remote = _env_bool("JOBSPY_IS_REMOTE", False)
term_index = _env_int("JOBSPY_TERM_INDEX", 1)
term_total = _env_int("JOBSPY_TERM_TOTAL", 1)
output_csv = Path(_env_str("JOBSPY_OUTPUT_CSV", "jobs.csv"))
output_json = Path(
@@ -50,6 +60,14 @@ def main() -> int:
output_json.parent.mkdir(parents=True, exist_ok=True)
print(f"jobspy: Search term: {search_term}")
_emit_progress(
"term_start",
{
"termIndex": term_index,
"termTotal": term_total,
"searchTerm": search_term,
},
)
jobs = scrape_jobs(
site_name=sites,
search_term=search_term,
@@ -62,6 +80,15 @@ def main() -> int:
)
print(f"Found {len(jobs)} jobs")
_emit_progress(
"term_complete",
{
"termIndex": term_index,
"termTotal": term_total,
"searchTerm": search_term,
"jobsFoundTerm": int(len(jobs)),
},
)
jobs.to_csv(
output_csv,