diff --git a/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/CrawlerRetreiver.java b/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/CrawlerRetreiver.java index 09a82367..457c524c 100644 --- a/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/CrawlerRetreiver.java +++ b/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/CrawlerRetreiver.java @@ -9,9 +9,6 @@ import nu.marginalia.crawl.retreival.fetcher.warc.WarcRecorder; import nu.marginalia.crawl.retreival.revisit.CrawlerRevisitor; import nu.marginalia.crawl.retreival.revisit.DocumentWithReference; import nu.marginalia.crawl.retreival.sitemap.SitemapFetcher; -import nu.marginalia.crawling.body.HttpFetchResult; -import nu.marginalia.crawling.model.CrawledDomain; -import nu.marginalia.crawling.model.CrawlerDomainStatus; import nu.marginalia.ip_blocklist.UrlBlocklist; import nu.marginalia.link_parser.LinkParser; import nu.marginalia.model.EdgeDomain; diff --git a/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcher.java b/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcher.java index d0a8b075..42723d5c 100644 --- a/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcher.java +++ b/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcher.java @@ -4,7 +4,6 @@ import com.google.inject.ImplementedBy; import crawlercommons.robots.SimpleRobotRules; import nu.marginalia.crawl.retreival.RateLimitException; import nu.marginalia.crawl.retreival.fetcher.warc.WarcRecorder; -import nu.marginalia.crawling.body.HttpFetchResult; import nu.marginalia.model.EdgeDomain; import nu.marginalia.model.EdgeUrl; import nu.marginalia.model.body.HttpFetchResult; diff --git a/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcherImpl.java b/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcherImpl.java index 49ac03bc..40725b0f 100644 --- a/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcherImpl.java +++ b/code/processes/crawling-process/java/nu/marginalia/crawl/retreival/fetcher/HttpFetcherImpl.java @@ -12,9 +12,6 @@ import nu.marginalia.crawl.retreival.fetcher.socket.FastTerminatingSocketFactory import nu.marginalia.crawl.retreival.fetcher.socket.IpInterceptingNetworkInterceptor; import nu.marginalia.crawl.retreival.fetcher.socket.NoSecuritySSL; import nu.marginalia.crawl.retreival.fetcher.warc.WarcRecorder; -import nu.marginalia.crawling.body.ContentTypeLogic; -import nu.marginalia.crawling.body.DocumentBodyExtractor; -import nu.marginalia.crawling.body.HttpFetchResult; import nu.marginalia.model.EdgeDomain; import nu.marginalia.model.EdgeUrl; import nu.marginalia.model.body.ContentTypeLogic; diff --git a/code/processes/crawling-process/model/java/nu/marginalia/io/crawldata/CrawledDomainReader.java b/code/processes/crawling-process/model/java/nu/marginalia/io/crawldata/CrawledDomainReader.java index 7588bbaa..7e359814 100644 --- a/code/processes/crawling-process/model/java/nu/marginalia/io/crawldata/CrawledDomainReader.java +++ b/code/processes/crawling-process/model/java/nu/marginalia/io/crawldata/CrawledDomainReader.java @@ -1,7 +1,6 @@ package nu.marginalia.io.crawldata; import nu.marginalia.io.crawldata.format.ParquetSerializableCrawlDataStream; -import nu.marginalia.crawling.io.format.ParquetSerializableCrawlDataStream; import org.slf4j.Logger; import org.slf4j.LoggerFactory; diff --git a/code/processes/crawling-process/test/nu/marginalia/crawling/HttpFetcherTest.java b/code/processes/crawling-process/test/nu/marginalia/crawling/HttpFetcherTest.java index af196da7..611cc8c2 100644 --- a/code/processes/crawling-process/test/nu/marginalia/crawling/HttpFetcherTest.java +++ b/code/processes/crawling-process/test/nu/marginalia/crawling/HttpFetcherTest.java @@ -6,10 +6,10 @@ import nu.marginalia.crawl.retreival.fetcher.ContentTags; import nu.marginalia.crawl.retreival.fetcher.HttpFetcher; import nu.marginalia.crawl.retreival.fetcher.HttpFetcherImpl; import nu.marginalia.crawl.retreival.fetcher.warc.WarcRecorder; -import nu.marginalia.crawling.body.ContentTypeLogic; -import nu.marginalia.crawling.body.DocumentBodyExtractor; -import nu.marginalia.crawling.body.DocumentBodyResult; import nu.marginalia.model.EdgeUrl; +import nu.marginalia.model.body.ContentTypeLogic; +import nu.marginalia.model.body.DocumentBodyExtractor; +import nu.marginalia.model.body.DocumentBodyResult; import org.junit.jupiter.api.Assertions; import org.junit.jupiter.api.Test;