mirror of
https://git.mirrors.martin98.com/https://github.com/mendableai/firecrawl
synced 2025-08-20 07:59:11 +08:00
Nick: we are using runpod
This commit is contained in:
parent
5fcf3fa97e
commit
b8d7f9f257
@ -13,12 +13,12 @@ import path from "node:path";
|
|||||||
|
|
||||||
type PDFProcessorResult = { html: string; markdown?: string };
|
type PDFProcessorResult = { html: string; markdown?: string };
|
||||||
|
|
||||||
async function scrapePDFWithMinerU(
|
async function scrapePDFWithRunPodMU(
|
||||||
meta: Meta,
|
meta: Meta,
|
||||||
tempFilePath: string,
|
tempFilePath: string,
|
||||||
timeToRun: number | undefined,
|
timeToRun: number | undefined,
|
||||||
): Promise<PDFProcessorResult> {
|
): Promise<PDFProcessorResult> {
|
||||||
meta.logger.debug("Processing PDF document with MinerU", {
|
meta.logger.debug("Processing PDF document with RunPod MU", {
|
||||||
tempFilePath,
|
tempFilePath,
|
||||||
});
|
});
|
||||||
|
|
||||||
@ -30,10 +30,10 @@ async function scrapePDFWithMinerU(
|
|||||||
console.log(tempFilePath);
|
console.log(tempFilePath);
|
||||||
|
|
||||||
const upload = await robustFetch({
|
const upload = await robustFetch({
|
||||||
url: "https://api.runpod.ai/v2/" + process.env.MINERU_POD_ID + "/run",
|
url: "https://api.runpod.ai/v2/" + process.env.RUNPOD_MU_POD_ID + "/run",
|
||||||
method: "POST",
|
method: "POST",
|
||||||
headers: {
|
headers: {
|
||||||
Authorization: `Bearer ${process.env.MINERU_API_KEY}`,
|
Authorization: `Bearer ${process.env.RUNPOD_MU_API_KEY}`,
|
||||||
},
|
},
|
||||||
body: {
|
body: {
|
||||||
input: {
|
input: {
|
||||||
@ -42,7 +42,7 @@ async function scrapePDFWithMinerU(
|
|||||||
},
|
},
|
||||||
},
|
},
|
||||||
logger: meta.logger.child({
|
logger: meta.logger.child({
|
||||||
method: "scrapePDFWithMinerU/upload/robustFetch",
|
method: "scrapePDFWithRunPodMU/upload/robustFetch",
|
||||||
}),
|
}),
|
||||||
schema: z.object({
|
schema: z.object({
|
||||||
id: z.string(),
|
id: z.string(),
|
||||||
@ -58,13 +58,13 @@ async function scrapePDFWithMinerU(
|
|||||||
while (Date.now() <= startedAt + timeout) {
|
while (Date.now() <= startedAt + timeout) {
|
||||||
try {
|
try {
|
||||||
const result = await robustFetch({
|
const result = await robustFetch({
|
||||||
url: `https://api.runpod.ai/v2/${process.env.MINERU_POD_ID}/status/${jobId}`,
|
url: `https://api.runpod.ai/v2/${process.env.RUNPOD_MU_POD_ID}/status/${jobId}`,
|
||||||
method: "GET",
|
method: "GET",
|
||||||
headers: {
|
headers: {
|
||||||
Authorization: `Bearer ${process.env.MINERU_API_KEY}`,
|
Authorization: `Bearer ${process.env.RUNPOD_MU_API_KEY}`,
|
||||||
},
|
},
|
||||||
logger: meta.logger.child({
|
logger: meta.logger.child({
|
||||||
method: "scrapePDFWithMinerU/result/robustFetch",
|
method: "scrapePDFWithRunPodMU/result/robustFetch",
|
||||||
}),
|
}),
|
||||||
schema: z.object({
|
schema: z.object({
|
||||||
status: z.string(),
|
status: z.string(),
|
||||||
@ -83,7 +83,7 @@ async function scrapePDFWithMinerU(
|
|||||||
}
|
}
|
||||||
|
|
||||||
if (result.status === "FAILED") {
|
if (result.status === "FAILED") {
|
||||||
throw new Error("MinerU failed to parse PDF: " + result.error!, { cause: result.error });
|
throw new Error("RunPod MU failed to parse PDF: " + result.error!, { cause: result.error });
|
||||||
}
|
}
|
||||||
|
|
||||||
// result not up yet
|
// result not up yet
|
||||||
@ -96,7 +96,7 @@ async function scrapePDFWithMinerU(
|
|||||||
// meta.logger.debug("URL is not actually a PDF, signalling...");
|
// meta.logger.debug("URL is not actually a PDF, signalling...");
|
||||||
// throw new RemoveFeatureError(["pdf"]);
|
// throw new RemoveFeatureError(["pdf"]);
|
||||||
// } else {
|
// } else {
|
||||||
throw new Error("MinerU threw an error", {
|
throw new Error("RunPod MU threw an error", {
|
||||||
cause: e.cause,
|
cause: e.cause,
|
||||||
});
|
});
|
||||||
// }
|
// }
|
||||||
@ -108,7 +108,7 @@ async function scrapePDFWithMinerU(
|
|||||||
await new Promise<void>((resolve) => setTimeout(() => resolve(), 250));
|
await new Promise<void>((resolve) => setTimeout(() => resolve(), 250));
|
||||||
}
|
}
|
||||||
|
|
||||||
throw new Error("MinerU timed out");
|
throw new Error("RunPod MU timed out");
|
||||||
}
|
}
|
||||||
|
|
||||||
async function scrapePDFWithParsePDF(
|
async function scrapePDFWithParsePDF(
|
||||||
@ -157,33 +157,33 @@ export async function scrapePDF(
|
|||||||
tempFilePath,
|
tempFilePath,
|
||||||
);
|
);
|
||||||
|
|
||||||
// Then, if output is too short, pass to MinerU
|
// Then, if output is too short, pass to RunPod MU
|
||||||
if (
|
if (
|
||||||
result.markdown && result.markdown.length < 500 &&
|
result.markdown && result.markdown.length < 500 &&
|
||||||
process.env.MINERU_API_KEY && process.env.MINERU_POD_ID
|
process.env.RUNPOD_MU_API_KEY && process.env.RUNPOD_MU_POD_ID
|
||||||
) {
|
) {
|
||||||
try {
|
try {
|
||||||
const mineruResult = await scrapePDFWithMinerU(
|
const muResult = await scrapePDFWithRunPodMU(
|
||||||
{
|
{
|
||||||
...meta,
|
...meta,
|
||||||
logger: meta.logger.child({
|
logger: meta.logger.child({
|
||||||
method: "scrapePDF/scrapePDFWithMinerU",
|
method: "scrapePDF/scrapePDFWithRunPodMU",
|
||||||
}),
|
}),
|
||||||
},
|
},
|
||||||
tempFilePath,
|
tempFilePath,
|
||||||
timeToRun,
|
timeToRun,
|
||||||
);
|
);
|
||||||
result = mineruResult; // Use LlamaParse result if successful
|
result = muResult; // Use LlamaParse result if successful
|
||||||
} catch (error) {
|
} catch (error) {
|
||||||
if (error instanceof Error && error.message === "MinerU timed out") {
|
if (error instanceof Error && error.message === "RunPod MU timed out") {
|
||||||
meta.logger.warn("MinerU timed out -- using parse-pdf result", {
|
meta.logger.warn("RunPod MU timed out -- using parse-pdf result", {
|
||||||
error,
|
error,
|
||||||
});
|
});
|
||||||
} else if (error instanceof RemoveFeatureError) {
|
} else if (error instanceof RemoveFeatureError) {
|
||||||
throw error;
|
throw error;
|
||||||
} else {
|
} else {
|
||||||
meta.logger.warn(
|
meta.logger.warn(
|
||||||
"MinerU failed to parse PDF -- using parse-pdf result",
|
"RunPod MU failed to parse PDF -- using parse-pdf result",
|
||||||
{ error },
|
{ error },
|
||||||
);
|
);
|
||||||
Sentry.captureException(error);
|
Sentry.captureException(error);
|
||||||
|
Loading…
x
Reference in New Issue
Block a user