Examples with CrawlWrapperException won.matcher.service.crawler.exception.CrawlWrapperException used on opensource projects

Example 1 with CrawlWrapperException

use of won.matcher.service.crawler.exception.CrawlWrapperException in project webofneeds by researchstudio-sat.

the class WorkerCrawlerActor method crawlUri.

private void crawlUri(CrawlUriMessage uriMsg) {
    Dataset ds = null;
    List<String> etags = null;
    Lock lock = null;
    try {
        // check if resource is already downloaded
        if (uriMsg instanceof ResourceCrawlUriMessage) {
            ResourceCrawlUriMessage resMsg = ((ResourceCrawlUriMessage) uriMsg);
            if (resMsg.getSerializedResource() != null && resMsg.getSerializationFormat() != null) {
                // TODO: this should be optimized, why deserialize the resource here when we
                // just want to save it in the RDF
                // store? How to insert this serialized resource into the SPARQL endpoint?
                ds = SparqlService.deserializeDataset(resMsg.getSerializedResource(), resMsg.getSerializationFormat());
            }
        }
        // download resource if not already downloaded
        if (ds == null) {
            // use ETag/If-None-Match Headers to make the process more efficient
            HttpHeaders httpHeaders = new HttpHeaders();
            if (uriMsg.getResourceETagHeaderValues() != null && !uriMsg.getResourceETagHeaderValues().isEmpty()) {
                String ifNoneMatchHeaderValue = StringUtils.collectionToDelimitedString(uriMsg.getResourceETagHeaderValues(), ", ");
                httpHeaders.add("If-None-Match", ifNoneMatchHeaderValue);
            }
            DatasetResponseWithStatusCodeAndHeaders datasetWithHeaders = linkedDataSource.getDatasetWithHeadersForResource(URI.create(uriMsg.getUri()), httpHeaders);
            ds = datasetWithHeaders.getDataset();
            etags = datasetWithHeaders.getResponseHeaders().get("ETag");
            // if dataset was not modified (304) we can treat the current crawl uri as done
            if (ds == null && datasetWithHeaders.getStatusCode() == 304) {
                sendDoneUriMessage(uriMsg, uriMsg.getWonNodeUri(), etags);
                return;
            }
            // if there is paging activated and the won node tells us that there is more
            // data (previous link)
            // to be downloaded, then we add this link to the crawling process too
            String prevLink = linkedDataSource.getPreviousLinkFromDatasetWithHeaders(datasetWithHeaders);
            if (prevLink != null) {
                CrawlUriMessage newUriMsg = new CrawlUriMessage(uriMsg.getBaseUri(), prevLink, uriMsg.getWonNodeUri(), CrawlUriMessage.STATUS.PROCESS, System.currentTimeMillis(), null);
                getSender().tell(newUriMsg, getSelf());
            }
        }
        lock = ds == null ? null : ds.getLock();
        lock.enterCriticalSection(true);
        // Save dataset to triple store
        sparqlService.updateNamedGraphsOfDataset(ds);
        String wonNodeUri = extractWonNodeUri(ds, uriMsg.getUri());
        if (wonNodeUri == null) {
            wonNodeUri = uriMsg.getWonNodeUri();
        }
        // do nothing more here if the STATUS of the message was SAVE
        if (uriMsg.getStatus().equals(CrawlUriMessage.STATUS.SAVE)) {
            log.debug("processed crawl uri event {} with status 'SAVE'", uriMsg);
            return;
        }
        // extract URIs from current resource and send extracted URI messages back to
        // sender
        log.debug("Extract URIs from message {}", uriMsg);
        Set<CrawlUriMessage> newCrawlMessages = sparqlService.extractCrawlUriMessages(uriMsg.getBaseUri(), wonNodeUri);
        for (CrawlUriMessage newMsg : newCrawlMessages) {
            getSender().tell(newMsg, getSelf());
        }
        // signal sender that this URI is processed and save meta data about crawling
        // the URI.
        // This needs to be done after all extracted URI messages have been sent to
        // guarantee consistency
        // in case of failure
        sendDoneUriMessage(uriMsg, wonNodeUri, etags);
        // bu
        if (AtomModelWrapper.isAAtom(ds)) {
            AtomModelWrapper atomModelWrapper = new AtomModelWrapper(ds, false);
            AtomState state = atomModelWrapper.getAtomState();
            AtomEvent.TYPE type = state.equals(AtomState.ACTIVE) ? AtomEvent.TYPE.ACTIVE : AtomEvent.TYPE.INACTIVE;
            log.debug("Created atom event for atom uri {}", uriMsg.getUri());
            long crawlDate = System.currentTimeMillis();
            AtomEvent atomEvent = new AtomEvent(uriMsg.getUri(), wonNodeUri, type, crawlDate, ds, Cause.CRAWLED);
            pubSubMediator.tell(new DistributedPubSubMediator.Publish(atomEvent.getClass().getName(), atomEvent), getSelf());
        }
    } catch (RestClientException e1) {
        // usually happens if the fetch of the dataset fails e.g.
        // HttpServerErrorException, HttpClientErrorException
        log.debug("Exception during crawling: " + e1);
        throw new CrawlWrapperException(e1, uriMsg);
    } catch (LinkedDataFetchingException e) {
        log.debug("Exception during crawling: " + e);
        Throwable cause = e.getCause();
        if (cause instanceof HttpClientErrorException && Objects.equals(((HttpClientErrorException) cause).getStatusCode(), HttpStatus.GONE)) {
            log.debug("Uri used to exist, but has been deleted, deleting from rdf store.");
            sendDeletedAtomMessage(uriMsg.getUri(), uriMsg.getWonNodeUri());
            sendDeletedUriMessage(uriMsg, uriMsg.getWonNodeUri(), etags);
        } else if (cause instanceof HttpClientErrorException && Objects.equals(((HttpClientErrorException) cause).getStatusCode(), HttpStatus.FORBIDDEN)) {
            log.debug("Not allowed to access uri, marking as done");
            sendDoneUriMessage(uriMsg, uriMsg.getWonNodeUri(), etags);
        } else {
            throw new CrawlWrapperException(e, uriMsg);
        }
    } catch (Exception e) {
        log.debug("Exception during crawling: " + e);
        throw new CrawlWrapperException(e, uriMsg);
    } finally {
        if (lock != null) {
            lock.leaveCriticalSection();
        }
    }
}