Text Analysis

How do I...

Extract All Plain Text
Extract Plain Text Without HTTP Headers
Extract Plain Text By Domain
Extract Plain Text by URL Pattern
Extract Plain Text Minus Boilerplate
Extract Plain Text Filtered by Date
Extract Plain Text Filtered by Language
Extract Plain Text Filtered by Keyword
Extract Raw HTML

For all the scripts below, you can type :paste into Spark Shell, paste the script, and then run it with CTRL+d:

Extract All Plain Text

Scala RDD

This script extracts the crawl date, domain, URL, and plain text from HTML files in the sample ARC data (and saves the output to out/). By default, HTTP headers are included in the plain text that is extracted.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(r.getContentString)))
  .saveAsTextFile("plain-text-rdd/")

Note that this will create a new directory to store the output, which cannot already exist.

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url"), $"url", removeHTML($"content"))
  .write.csv("plain-text-df/")

Python DF

from aut import *

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_html("content").alias("content"))
  .write.csv("plain-text-df/")

Extract Plain Text Without HTTP Headers

Scala RDD

If you want to remove HTTP headers, you can add one more command: RemoveHttpHeader. The script would then look like:

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-noheaders-rdd/")

As most plain text use cases do not require HTTP headers to be in the output, we are removing headers in the following examples.

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select(removeHTML(removeHTTPHeader($"content")))
  .write.csv("plain-text-noheaders-df/")

Python DF

from aut import *

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select(remove_html(remove_http_header("content")).alias("content"))\
  .write.csv("plain-text-noheaders-df/")

Extract Plain Text By Domain

Scala RDD

The following Spark script generates plain text renderings for all the web pages in a collection with a URL matching a filter string. In the example case, it will go through the collection and find all of the URLs within the "archive.org" domain.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepDomains(Set("www.archive.org"))
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-domain-rdd/")

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val domains = Array("www.archive.org")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url").alias("domains"), $"url", removeHTML(removeHTTPHeader($"content").alias("content")))
  .filter(hasDomains($"domain", lit(domains)))
  .write.csv("plain-text-domain-df/")

Python DF

from aut import *
from pyspark.sql.functions import col

domains = ["www.archive.org"]

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_html(remove_http_header("content")).alias("content"))\
  .filter(col("domain").isin(domains))\
  .write.csv("plain-text-domain-df/")

Extract Plain Text by URL Pattern

Scala RDD

The following Spark script generates plain text renderings for all the web pages in a collection with a URL matching a regular expression pattern. In the example case, it will go through a WARC file and find all of the URLs beginning with http://archive.org/details/, and save the text of those URLs.

The (?i) makes this query case insensitive.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepUrlPatterns(Set("(?i)http://www.archive.org/details/.*".r))
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("details-rdd/")

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val urlPattern = Array("(?i)http://www.archive.org/details/.*")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url").alias("domain"), $"url", removeHTML(removeHTTPHeader($"content").alias("content")))
  .filter(hasUrlPatterns($"url", lit(urlsPattern)))
  .write.csv("details-df/")

Python DF

from aut import *
from pyspark.sql.functions import col

url_pattern = "%http://www.archive.org/details/%"

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_html(remove_http_header("content")).alias("content"))\
  .filter(col("url").like(url_pattern))\
  .write.csv("details-df/")

Extract Plain Text Minus Boilerplate

Scala RDD

The following Spark script generates plain text renderings for all the web pages in a collection, minus "boilerplate" content: advertisements, navigational elements, and elements of the website template. For more information on the boilerplate removal library we are using, please see this website and paper.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepDomains(Set("www.archive.org"))
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, ExtractBoilerpipeTextRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-no-boilerplate-rdd/")

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val domains = Array("www.archive.org")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url"), $"url", extractBoilerpipeText(removeHTTPHeader($"content")))
  .filter(hasDomains($"domain", lit(domains)))
  .write.csv("plain-text-no-boilerplate-df/")

Python DF

from aut import *

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", extract_boilerplate(remove_http_header("content")).alias("content"))\
  .write.csv("plain-text-no-boilerplate-df/")

Extract Plain Text Filtered by Date

Scala RDD

AUT permits you to filter records by a list of full or partial date strings. It conceives of the date string as a DateComponent. Use keepDate to specify the year (YYYY), month (MM), day (DD), year and month (YYYYMM), or a particular year-month-day (YYYYMMDD).

The following Spark script extracts plain text for a given collection by date (in this case, April 2008).

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepDate(List("200804"), ExtractDateRDD.DateComponent.YYYYMM)
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-date-filtered-200804/")

The following script extracts plain text for a given collection by year (in this case, 2008).

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepDate(List("2008"), ExtractDateRDD.DateComponent.YYYY)
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-date-filtered-2008/")

Finally, you can also extract multiple dates or years. In this case, we would extract pages from both 2008 and 2015.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepDate(List("2008","2015"), ExtractDateRDD.DateComponent.YYYY)
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-date-filtered-2008-2015-rdd/")

Note: if you created a dump of plain text using another one of the earlier commands, you do not need to go back and run this. You can instead use bash to extract a sample of text. For example, running this command on a dump of all plain text stored in alberta_education_curriculum.txt:

sed -n -e '/^(201204/p' alberta_education_curriculum.txt > alberta_education_curriculum-201204.txt

would select just the lines beginning with (201204, or April 2012.

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val dates = Array("2008", "2015")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url").as("domain"), $"url", removeHTML(removeHTTPHeader($"content")).as("content"))
  .filter(hasDate($"crawl_date", lit(dates)))
  .write.csv("plain-text-date-filtered-2008-2015-df/")

Python DF

from aut import *
from pyspark.sql.functions import col

dates = "2009[10][09]\d\d"

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_html(remove_http_header("content")).alias("content"))\
  .filter(col("crawl_date").rlike(dates))\
  .write.csv("plain-text-date-filtered-2008-2015-df/")

Extract Plain Text Filtered by Language

Scala RDD

The following Spark script keeps only French language pages from a certain top-level domain. It uses the ISO 639.2 language codes.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .keepDomains(Set("www.archive.org"))
  .keepLanguages(Set("fr"))
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-fr-rdd/")

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val domains = Array("www.archive.org")
val languages = Array("fr")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url").alias("domain"), $"url", $"language", removeHTML(removeHTTPHeader($"content").alias("content")))
  .filter(hasDomains($"domain", lit(domains)))
  .filter(hasLanguages($"language", lit(languages)))
  .write.csv("plain-text-fr-df/")

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val domains = Array("www.archive.org")
val languages = Array("fr")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .filter(hasDomains(extractDomain($"url"), lit(domains)))
  .filter(hasLanguages($"language", lit(languages)))
  .select($"crawl_date", extractDomain($"url").alias("domain"), $"url", $"language", removeHTML(removeHTTPHeader($"content").alias("content")))
  .write.csv("plain-text-fr-df/")

Python DF

from aut import *
from pyspark.sql.functions import col

domains = ["www.geocities.com"]
languages = ["fr"]

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_html(remove_http_header("content")).alias("content"))\
  .filter(col("domain").isin(domains))\
  .filter(col("language").isin(languages))\
  .write.csv("plain-text-fr-df/")

Extract Plain text Filtered by Keyword

Scala RDD

The following Spark script keeps only pages containing a certain keyword, which also stacks on the other scripts.

For example, the following script takes all pages containing the keyword "radio" in a collection.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs",sc)
  .keepValidPages()
  .keepContent(Set("radio".r))
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTMLRDD(RemoveHTTPHeaderRDD(r.getContentString))))
  .saveAsTextFile("plain-text-radio-rdd/")

There is also discardContent which does the opposite, and can be used in cases where, for example, you have a frequent keyword you are not interested in.

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

val content = Array("radio")

RecordLoader.loadArchives("/path/to/warcs", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url").alias("domain"), $"url", removeHTML(removeHTTPHeader($"content").alias("content")))
  .filter(hasContent($"content", lit(content)))
  .write.csv("plain-text-radio-df/")

Python DF

from aut import *
from pyspark.sql.functions import col

content = "%radio%"

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_html(remove_http_header("content")).alias("content"))\
  .filter(col("content").like(content))
  .write.csv("plain-text-radio-df/")

Extract Raw HTML

Scala RDD

In most cases, users will be interested in working with plain text. In some cases, however, you may want to work with the actual HTML of the pages themselves (for example, looking for specific tags or HTML content).

The following script will produce the raw HTML of a WARC file. You can use the filters from above to filter it down accordingly by domain, language, etc.

import io.archivesunleashed._
import io.archivesunleashed.matchbox._

RecordLoader.loadArchives("/path/to/warcs", sc)
  .keepValidPages()
  .map(r => (r.getCrawlDate, r.getDomain, r.getUrl, RemoveHTTPHeaderRDD(r.getContentString)))
  .saveAsTextFile("plain-html-rdd/")

Scala DF

import io.archivesunleashed._
import io.archivesunleashed.udfs._

RecordLoader.loadArchives("example.warc.gz", sc)
  .webpages()
  .select($"crawl_date", extractDomain($"url"), $"url", removeHTTPHeader($"content"))
  .write.csv("plain-html-df/")

Python DF

from aut import *

WebArchive(sc, sqlContext, "/path/to/warcs")\
  .webpages()\
  .select("crawl_date", extract_domain("url").alias("domain"), "url", remove_http_header("content").alias("content"))\
  .write.csv("plain-html-df/")

Files

text-analysis.md

Latest commit

History

text-analysis.md

File metadata and controls

Text Analysis

Extract All Plain Text

Scala RDD

Scala DF

Python DF

Extract Plain Text Without HTTP Headers

Scala RDD

Scala DF

Python DF

Extract Plain Text By Domain

Scala RDD

Scala DF

Python DF

Extract Plain Text by URL Pattern

Scala RDD

Scala DF

Python DF

Extract Plain Text Minus Boilerplate

Scala RDD

Scala DF

Python DF

Extract Plain Text Filtered by Date

Scala RDD

Scala DF

Python DF

Extract Plain Text Filtered by Language

Scala RDD

Scala DF

Python DF

Extract Plain text Filtered by Keyword

Scala RDD

Scala DF

Python DF

Extract Raw HTML

Scala RDD

Scala DF

Python DF