Language identification may show poor performance on documents with short or no
title but clear lang indication in text content. Using content text too
improves lang detection.
+ remove double caching of text in Identificator
to support the new time parser and search functions in YaCy a high
precision detection of date and time on the day is necessary. That
requires that the time zone of the document content and the time zone of
the user, doing a search, is detected. The time zone of the search
request is done automatically using the browsers time zone offset which
is delivered to the search request automatically and invisible to the
user. The time zone for the content of web pages cannot be detected
automatically and must be an attribute of crawl starts. The advanced
crawl start now provides an input field to set the time zone in minutes
as an offset number. All parsers must get a time zone offset passed, so
this required the change of the parser java api. A lot of other changes
had been made which corrects the wrong handling of dates in YaCy which
was to add a correction based on the time zone of the server. Now no
correction is added and all dates in YaCy are UTC/GMT time zone, a
normalized time zone for all peers.
- date navigation
The date is taken from the CONTENT of the documents / web pages, NOT
from a date submitted in the context of metadata (i.e. http header or
html head form). This makes it possible to search for documents in the
future, i.e. when documents contain event descriptions for future
events.
The date is written to an index field which is now enabled by default.
All documents are scanned for contained date mentions.
To visualize the dates for a specific search results, a histogram
showing the number of documents for each day is displayed. To render
these histograms the morris.js library is used. Morris.js requires also
raphael.js which is now also integrated in YaCy.
The histogram is now also displayed in the index browser by default.
To select a specific range from a search result, the following modifiers
had been introduced:
from:<date>
to:<date>
These modifiers can be used separately (i.e. only 'from' or only 'to')
to describe an open interval or combined to have a closed interval. Both
dates are inclusive. To select a specific single date only, use the
'to:' - modifier.
The histogram shows blue and green lines; the green lines denot weekend
days (saturday and sunday).
Clicking on bars in the histogram has the following reaction:
1st click: add a from:<date> modifier for the date of the bar
2nd click: add a to:<date> modifier for the date of the bar
3rd click: remove from and date modifier and set a on:<date> for the bar
When the on:<date> modifier is used, the histogram shows an unlimited
time period. This makes it possible to click again (4th click) which is
then interpreted as a 1st click again (sets a from modifier).
The display feature is NOT switched on by default; to switch it on use
the /ConfigSearchPage_p.html servlet.
given css class and extends a given vocabulary with a term consisting
with the text content of the html class tag. Additionally, the term is
included into the semantic facet of the document. This allows the
creation of faceted search to documents without the pre-creation of
vocabularies; instead, the vocabulary is created on-the-fly, possibly
for use in other crawls. If any of the term scraping for a specific
vocabulary is successful on a document, this vocabulary is excluded for
auto-annotation on the page.
To use this feature, do the following:
- create a vocabulary on /Vocabulary_p.html (if not existent)
- in /CrawlStartExpert.html you will now see the vocabularies as column
in a table. The second column provides text fields where you can name
the class of html entities where the literal of the corresponding
vocabulary shall be scraped out
- when doing a search, you will see the content of the scraped fields in
a navigation facet for the given vocabulary
notions within the fulltext of a document. This class attempts to
identify also dates given abbreviated or with missing year or described
with names for special days, like 'Halloween'. In case that a date has
no year given, the current year and following years are considered.
This process is therefore able to identify a large set of dates to a
document, either because there are several dates given in the document
or the date is ambiguous. Four new Solr fields are used to store the
parsing result:
dates_in_content_sxt:
if date expressions can be found in the content, these dates are listed
here in order of the appearances
dates_in_content_count_i:
the number of entries in dates_in_content_sxt
date_in_content_min_dt:
if dates_in_content_sxt is filled, this contains the oldest date from
the list of available dates
#date_in_content_max_dt:
if dates_in_content_sxt is filled, this contains the youngest date from
the list of available dates, that may also be possibly in the future
These fields are deactiviated by default because the evaluation of
regular expressions to detect the date is yet too CPU intensive. Maybe
future enhancements will cause that this is switched on by default.
The purpose of these fields is the creation of calendar-like search
facets, to be implemented next.
the right content domain (i.e. identifying that it is an image, text
etc.) because it used the file extension and not an existing mime type
assignment.
- fixed the new setting that images shall be loaded for a better image
search.
- both fixes together makes it now possible to crawl
commons.wikimedia.org which makes use of 'funny' document names (i.e.
ending with .jpg while the document is html)
for anchor attributes.
- this caused that large portions of the parser code had to be adopted
as well
- added a counter target_order_i for anchor links in webgraph
computation
all unique links! This made it necessary, that a large portion of the
parser and link processing classes must be adopted to carry a different
type of link collection which carry a property attribute which are
attached to web anchors.
- introduction of a new URL class, AnchorURL
- the other url classes, DigestURI and MultiProtocolURI had been renamed
and refactored to fit into a new document package schema, document.id
- cleanup of net.yacy.cora.document package and refactoring
jdk-based logger tend to block
at java.util.logging.Logger.log(Logger.java:476) in concurrent
environments. This makes logging a main performance issue. To overcome
this problem, this is a add-on to jdk logging to put log entries on a
concurrent message queue and log the messages one by one using a
separate process.
- FTPClient uses the concurrent logging instead of the log4j logger
- an existing ranking servlet for solr was extended. It is now possible
to set boost values for fields, boost functions and boost queries.
- The ranking can have different instances, but currently only the first
one is used
- added an abstraction layer for fields which can be used for search and
those fields can be edited in the solr ranking configruation
- the ranking value from solr within the field score is used to combine
remote search requests, which all are created using the same locally
defined boost values
- reduced the number of fields which are used for search (makes it
faster)
- replaced some text fields by string fields (makes indexing faster)
- removed classes which had no use
- made a large number of experiments for a better ranking and created a
temporary setting which prefers hits inside titles
- adjusted also the RWI-based ranking computation to 'prefer title'
- made special cases like for portal search where no post-processing and
post-ranking is wanted: this keeps the original ranking order as done by
Solr
- fixed many bugs with old settings for ranking
The default schema uses only some of them and the resting search index
has now the following properties:
- webgraph size will have about 40 times as much entries as default
index
- the complete index size will increase and may be about the double size
of current amount
As testing showed, not much indexing performance is lost. The default
index will be smaller (moved fields out of it); thus searching
can be faster.
The new index will cause that some old parts in YaCy can be removed,
i.e. specialized webgraph data and the noload crawler. The new index
will make it possible to:
- search within link texts of linked but not indexed documents (about 20
times of document index in size!!)
- get a very detailed link graph
- enhance ranking using a complete link graph
To get the full access to the new index, the API to solr has now two
access points: one with attribute core=collection1 for the default
search index and core=webgraph to the new webgraph search index. This is
also avaiable for p2p operation but client access is not yet
implemented.
- migrates all entries in old urldb
Metadata coordinate (lat / lon) NumberFormatException still relative often (see excerpt below),
- added try/catch for URIMetadataRow (seems not to be needed in URIMetaDataNode, as Solr internally checks for number format)
- removed possible typ conversion for lat() / lon() comparison with 0.0f, changed to 0.0 (leaving it to the compiler/optimizer to choose number format)
current log excerpt for NumberFormatException:
W 2013/01/14 00:10:07 StackTrace For input string: "-"
java.lang.NumberFormatException: For input string: "-"
at sun.misc.FloatingDecimal.readJavaFormatString(Unknown Source)
at java.lang.Double.parseDouble(Unknown Source)
at net.yacy.kelondro.data.meta.URIMetadataRow$Components.lon(URIMetadataRow.java:525)
at net.yacy.kelondro.data.meta.URIMetadataRow.lon(URIMetadataRow.java:279)
at net.yacy.search.index.SolrConfiguration.metadata2solr(SolrConfiguration.java:277)
at net.yacy.search.index.Fulltext.putMetadata(Fulltext.java:329)
at transferURL.respond(transferURL.java:152)
...
Caused by: java.lang.NumberFormatException: For input string: "-"
at sun.misc.FloatingDecimal.readJavaFormatString(Unknown Source)
at java.lang.Double.parseDouble(Unknown Source)
at net.yacy.kelondro.data.meta.URIMetadataRow$Components.lon(URIMetadataRow.java:525)
at net.yacy.kelondro.data.meta.URIMetadataRow.lon(URIMetadataRow.java:279)
at net.yacy.search.index.SolrConfiguration.metadata2solr(SolrConfiguration.java:277)
at net.yacy.search.index.Fulltext.putMetadata(Fulltext.java:329)
at transferURL.respond(transferURL.java:152)
This uses an enhanced version of the Nutch/Solr TextProfileSignatue.
As a result, a signature of the document is written to the solr search
index. Additionally for each time when a signature is written, it is
checked if the singature exists already in the index. If the signature
does not exist, the document is marked as unique. The unique attribute
can now be used to sort document lists and bring duplicates to the end
of a result list.
To enable this, a large portion of the search api to Solr had to be
changed. This affected mainly caching of 'exists' searches to enhance
the check for existing signatures and do this without actually doing a
solr query.
Because here the first time a long number is used as value in the Solr
store, also the value naming in the YaCySchema had to be adopted and
normalized. This caused that many files had to be changed.
MultiProtocolURI during normalform computation because that should
always be done and also be done during initialization of the
MultiProtocolURI Object. The new normalform method takes only one
argument which should be 'true' unless you know exactly what you are
doing.
This can be used to add another stemming to solr using stemming files
that are expressed as synonyms for grammatical alternatives. The
synonym/stemming files must have the following form:
- each line is a comma-separated list of synonyms
- the list of synonyms may be enclosed with {} (like the GSA synonyms
file)
- the file may contain comments which are lines starting with a '#'
The synonym file(s) must be placed in DATA/DICTIONARIES/synonyms/ and
are activated by default whenever a synonym file is in place.
Then, for each word that is found in a document all synonyms are added
to a long text field which is stored into synonyms_t. Processes using
the synonyms must query with that field as optional matcher.
writings to the Metadata-DB are now also done to solr. This includes
metadata transfer during search and rwi transfer.
The new/added solr fields are:
## time when resource was loaded
load_date_dt
## date until resource shall be considered as fresh
fresh_date_dt
## id of the host, a 6-byte hash that is part of the document id
host_id_s
## ids of referrer to this document
referrer_id_ss
## the md5 of the raw source
md5_s
## the name of the publisher of the document
publisher_t
## the language used in the document; starts with primary language
language_ss
## an external ranking value
ranking_i
## the size of the raw source
size_i
## number of links to audio resources
audiolinkscount_i
## number of links to video resources
videolinkscount_i
## number of links to application resources
applinkscount_i
of the major CPU users during snippet verification. The class was not
efficient for two reasons:
- it used a too complex input stream; generated from sources and UTF8
byte-conversions. The BufferedReader applied a strong overhead.
- to feed data into the SentenceReader, multiple toString/getBytes had
been applied until a buffered Reader from an input stream was possible.
These superfluous conversions had been removed.
- the best source for the Sentence Reader is a String. Therefore the
production of Strings had been forced inside the Document class.
- vocabulary annotation is not done any more into the metadata of urldb
- vocabularies are written into the jena triplestore using a rdf
vocabulary
- vocabularies for rdf tripel must be updated; refactoring done
- with the new navigation tags in the triplestore a faster
pre-urldb-lookup is possible: navigation is processed now within the RWI
during pre-ranking retrieval
- added also a Owl vocabulary stub to add the plain-text url to the
triplestore using the owl:sameas predicate
later
- added abstract add, delete, get methods in the triplestore
- added generation of triples after auto-annotation
- migrated all MultiProtocolURI objects to DigestURI in the parser since
the url hash is needed as subject value in the triples in the triple
store
- added log warnings in case that search processes run into time-out
situations
- better concurrency for Integer formatter (used a non-synchronized
formatter before)
- bugfix for search termination (a poison pill was missing)
- added timeout parameters for search (again) -> target is, that they
are never reached.
only links where the content can be parsed. All non-parseable links are
placed into the noload queue. The search process must therefore be able
to filter out non-text search results.
- This fixes the problem that image search results appeared in the text
search.
- The interactive search can retrieve now ALL types of links
- The p2p interface is now extended to retrieve only certain types of
links (text, image, video, apps)
- The search process has an extension to filter the right document type
according to the search query
- tags that are automatically generated now enrich the dc:subject
- auto-generated tags have a '$' at the beginning of the tag
- auto-generated tags lead the tag name with a vocabulary name
each tag has the form
$<vocabulary-name>:<tag-printname-space-replaced-by-'_'>
used a ASCII String <-> byte[] conversion wherever possible. Many Strings in YaCy are hashes which are pure ASCII (base64 hashes).
The new ASCII String <-> byte[] conversion method have less computation overhead than the UTF8 conversion.
git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@7746 6c8d7289-2bf4-0310-a012-ef5d649a1542
luccioman