yujunjun/firecrawl
1
0
Fork 0
mirror of https://github.com/mendableai/firecrawl.git synced 2025-11-23 13:41:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

Revert "feat(queue-worker): always crawl links from content even if sitemapped"

Browse Source 
This reverts commit 3c045c43a446bb7895892338c881cd7bc4f77cbf.
This commit is contained in:
Nicolas 2024-10-01 14:20:23 -03:00
parent 445fc432e9
commit 258c67ce67
1 changed files with 39 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
apps/api/src/services/queue-worker.ts
View File

@ -300,6 +300,7 @@ async function processJob(job: Job, token: string) {
const sc = (await getCrawl(job.data.crawl_id)) as StoredCrawl; const sc = (await getCrawl(job.data.crawl_id)) as StoredCrawl;
if (!job.data.sitemapped) {
if (!sc.cancelled) { if (!sc.cancelled) {
const crawler = crawlToCrawler(job.data.crawl_id, sc); const crawler = crawlToCrawler(job.data.crawl_id, sc);
@ -311,6 +312,7 @@ async function processJob(job: Job, token: string) {
for (const link of links) { for (const link of links) {
if (await lockURL(job.data.crawl_id, sc, link)) { if (await lockURL(job.data.crawl_id, sc, link)) {
// This seems to work really welel
const jobPriority = await getJobPriority({ const jobPriority = await getJobPriority({
plan: sc.plan as PlanType, plan: sc.plan as PlanType,
team_id: sc.team_id, team_id: sc.team_id,
@ -318,6 +320,11 @@ async function processJob(job: Job, token: string) {
}); });
const jobId = uuidv4(); const jobId = uuidv4();
// console.log("plan: ", sc.plan);
// console.log("team_id: ", sc.team_id)
// console.log("base priority: ", job.data.crawl_id ? 20 : 10)
// console.log("job priority: " , jobPriority, "\n\n\n")
const newJob = await addScrapeJob( const newJob = await addScrapeJob(
{ {
url: link, url: link,
@ -338,6 +345,7 @@ async function processJob(job: Job, token: string) {
} }
} }
} }
}
if (await finishCrawl(job.data.crawl_id)) { if (await finishCrawl(job.data.crawl_id)) {