moved and was not cleared anymore. This results in an huge fieldcache.
(http://lucene.apache.org/#highlights-of-the-lucene-release-includehttps://issues.apache.org/jira/browse/LUCENE-5666)
Here I try to use DovValues where it is possible.
For this I used the Api-Scheme as new basis für the Solr-Schema.
This needs at least a complete optimization of the Solr-Index to get a
smaller FieldCache.
Everything that is indexed with these setting will not use the
Fieldcache at all.
bayesian filters. This can be used to classify documents during
indexing-time using a pre-definied bayesian filter.
New wordings:
- a context is a class where different categories are possible. The
context name is equal to a facet name.
- a category is a facet type within a facet navigation. Each context
must have several categories, at least one custom name (things you want
to discover) and one with the exact name "negative".
To use this, you must do:
- for each context, you must create a directory within
DATA/CLASSIFICATION with the name of the context (the facet name)
- within each context directory, you must create text files with one
document each per line for every categroy. One of these categories MUST
have the name 'negative.txt'.
Then, each new document is classified to match within one of the given
categories for each context.
- fixed superfluous space in query field list
- fixed filter query logic
- removed look-ahead query which caused that each new search page
submitted two solr queries
- fixed random solr result orders in case that the solr score was equal:
this was then re-ordered by YaCy using the document hash which came from
the solr object and that appeared to be random. Now the hash of the url
is used and the score is additionally modified by the url length to
prevent that this particular case appears at all.
Background: some user report problem with connecting/crawling some sites via https which require SNI support (by default switched off in YaCy). On the other hand systems not demanding SNI support are sometimes not properly configured and due to a bug/feature in java 1.7 connection is aborted. The later is more often the case, so the default is still fine. With the java start parameter expert user can no alter the startparameter to -Djsse.enableSNIExtension=true (java default) if they crawl more hosts requiring SNI support.
The alternative to let YaCy try both during https handshake (deep inside the httpclient) is not pursut at this time.
during surrogate reading: those attributes from the dump are removed
during the import process and replaced by new detected attributes
according to the setting of the YaCy peer.
This may cause that all such attributes are removed if the importing
peer has no synonyms and/or no vocabularies defined.
index has more than two million documents. This protects the index from
beeing flooded with search requests that cannot be resolved before the
real search query has to be computet.
This is a very complex migration: many classes had been renamed or
removed, dependencies changed and the solr index type is now aligned to
be a solr cloud repository.
Together with the Solr 5.2 library update, one other dependent library
had been updated as well: httpclient 4.4->4.4.1
Older indexes are migrated from 4_10 to 5_2. However, the new index
structure is more efficient and we recommend to re-index everything.
Please use the index export before you do the update to a large
surrogate xml file. After the update, start with an empty index and then
initialize this with your dump.
This was caused by reading via GZIP from close-wait connection an caused
high cpu- and system-loads.
Instat of implementing handling of the RedListener now I found a
timelimeted 'get' "realy" solving this problem.
This was caused by reading via GZIP from close-wait connection an caused
high cpu- and system-loads.
Solved by implementing handling of the RedListener.
reason: experimental implementatin of RDFa parser not executed (limited to special urls) but may cause error on normal html parsing due to a inputstream.reset
export function is also now the default export option. The export file
format for a full solr export is very similar to a solr search result
xml, only the <lst name="responseHeader"> tag is missing.
The exported xml has a special line termination feature: all documents
will be exported into a single line without any CR in between. That
means that every document is completely inside a single line. While this
is not readable at all for humans, it is very useful for linux line
processing scripts, like grep. Using grep it will be easy to select
single documents which match for a given pattern.
Such dumps shall be importable with the DATA/SURROGATE/in import
function, but that import is not yet adopted to the new file format.
by instatiation from URIMetadataNode, by eleminating differentiation of ResultEntry/URIMetadataNode.
- moved remaining ResultEntry functionallity to URIMetadataNode
- for 1:1 functionallity added a function makeResultEntry()
- removed ResultEntry
- refactored related code
Main difference is after makeResultEntry the text_t content is removed and alternative title/url strings for display are calculated.
Main difference left is, that
If remote crawl option is not activated, skip init of remoteCrawlJob to save the resources of queue and ideling thread.
Deploy of the remoteCrawlJob deferred on activation of the option.
added to IndexReIndexMonitor_p.html
Selects existing documents from index and feeds it to the crawler.
currently only the field fresh_date_dt is used determine documents for recrawl (fresh_date_dt:[* TO NOW-1DAY]
Documents are added in small chunks (200) to the crawler, only if no other crawl is running.
keeping surrogates after processing is essential for some users. If the
space they are taking is too high, please set up an automatic deletion
process (like a cronjob).
This assigns priorities to incoming requests. Higher priority numbers are served before lower.
(disabled by default in defaults/web.xml,
uncomment or copy entry to DATA/Settings/web.xml)
as we flag if a peer is accesible via https, we need to know the port if we want to use is (e.g. for interYaCy communication)
start to provide / tansport the port by recording it in peers dna.
- add https link on the Network.html lock symbol
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.
collections in search result. When selecting one of them in another
search, switch off the previously selected collection. This actually
turns the collection navigation modifier into a radio-button like
behaviour
- if less or equal of 8 facet options are present, they are shown by
default
- if more facet options are present, they are hidden
To view or hide all facets, just click on the facet header bar
for location nav facet of field coordinate_p does not return results, now using coordinate_p_0_coordinate as alternative to get facet counts. As the actual facet value is not used this should not harm any analysis (even if facet is a incomplete location).
If facet value is used in future likely *_geohash field could be introduced (for facet and other ... as transport value)
interface to distinguish rich and poor document data.
This also reverts some changes from commit
796770e070 because the firstSeen database
is the wrong method to distinguish these types of data
To protect rich index data (full resource) from overwriting by metadata gathered during remote search,
the newly introduced "firstSeen" index is used to differentiate between full-resource-doc and metadata,
as a "firstSeen" entry is only added on store's of full-resource-docs (during crawl or remote search).
So far we used same escape procedure for all parts of the url (which includes x-www-form-urlencoded for all url components)
Added capability to use different encoding rules for the different url components (through specific bitset for each component).
(this is inspired by org.apache.http.client and java.net.uri implementation).
- Added test case for http://mantis.tokeek.de/view.php?id=559
http://mantis.tokeek.de/view.php?id=505
It happens but not able to reproduce. This change makes sure terminate signal is catched at end of currently running merge jobs
to considere description_txt always (solr hl & internal).
For now just added desc to text list for computation, could be further equalized with hl computation.
- if an eventDate is given in the search result, replace the document
date with the event date and prefix it with the string "on ".
- the document date is omitted if a date from the cent is shown
Added also the date as fields in the json and rss result sets.
- 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.
calculated from boostfields config, making sure title, description, keywords and content is always searched.
- apply change to solrServlet makes sure every remote query uses at least all locally defined boost fields for search
- apply to local solr search
- simplify select query by using QF defaults
If the url contains special chars (like umlaute äöü) it's interpreted as multybyte char and actually not converted at all (removed).
Added a check if the multibyte convesion is not complete, just add the char as is.
This fixes http://mantis.tokeek.de/view.php?id=200
This is useful in intranet indexing while crawling a intranet file server accessed via hostname while e.g. under Windows mapped to different drive letters on individual clients.
Here you can crawl e.g. file://fileserver/documents having a valid uri in that intranet environment (while e.g. P:/documents might be client dependant).
After several tests found that eom is not prevented. Major reason in testing was assumption future GC will free avg of last 5 GC.
Disabeling this check improved eom exceptions.
Added simplest testcase used for verification
did not terminate:
Occurrences: 100
at java.util.HashMap$KeyIterator.next(HashMap.java:956)
at
net.yacy.cora.protocol.ConnectionInfo.cleanup(ConnectionInfo.java:300)
at
net.yacy.cora.protocol.ConnectionInfo.cleanUp(ConnectionInfo.java:293)
at net.yacy.search.Switchboard.cleanupJob(Switchboard.java:2212)
at sun.reflect.GeneratedMethodAccessor12.invoke(Unknown Source)
at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:606)
at
net.yacy.kelondro.workflow.InstantBusyThread.job(InstantBusyThread.java:105)
at
net.yacy.kelondro.workflow.AbstractBusyThread.run(AbstractBusyThread.java:215)
This tries to fix the problem; the problem should be monitored
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
trace: java.lang.OutOfMemoryError: Java heap space
at java.awt.image.DataBufferInt.<init>(DataBufferInt.java:75)
at java.awt.image.Raster.createPackedRaster(Raster.java:467)
at java.awt.image.DirectColorModel.createCompatibleWritableRaster(DirectColorModel.java:1032)
at java.awt.image.BufferedImage.<init>(BufferedImage.java:331)
at net.yacy.document.parser.images.bmpParser$IMAGEMAP.<init>(bmpParser.java:149)
at net.yacy.document.parser.images.bmpParser.parse(bmpParser.java:69)
at net.yacy.document.parser.images.genericImageParser.parse(genericImageParser.java:116)
script. Adopted the YaCyApp environment and fixed a problem in the
startYACY.sh application wrapper which caused wrong usage of logging
option -l which caused that files had been written to the YaCy
application folder.
As a result of this fix, it is not necessary any more to change path
settings in Info.plist if libraries are changed.
new users. Will be activated maybe again if there is an optional
tutorial mode which can be switched on for this special purpose of
running a tutorial.
introduce FederateSearchManager handling search heuristic to external systems via specific FederateSearchConnectors,
which provide the query() functionallity, the translation to YaCy schema .toYaCySchema() and the search() routine to deliver results to searchevents, which is generally implemented in Abstract connector.
The manager enforces now a min 15s delay between calls to external systems.
Besides the OpensearchConnector a SolrFederateSearchConnector is available. It uses a additional config file for fieldname translation.
default heuristicopensearch.conf:
- openbdb.com removed - seems not longer to deliver results
- config via solrconnector to datacite.org added (large technical library archive)
Intended for searches/research projects with not sufficient results from local and DHT selected remote target peers.
Function: the process checks newly created bookmarks for description starting with "query=..." and takes this to ask every peer for 20 search results and adds it to the local index in a background job.
link to start/stop the process added to /Bookmarks.html
This is intended for peers who want to participate in the P2P network but don't wish to load/fill-up their index with metadata of every received search result.
The DHT transfer is not effected by this option (and will work as usual, so that a peer disabling the new store to index switch still receives and holds the metadata according to DHT rules).
Downside for the local peer is that search speed will not improve if search terms are only avail. remote or by quick hits in local index.
To be able to improve the local index a Click-Servlet option was added additionally.
If switched on, all search result links point to this servlet, which forwards the users browser (by html header) to the desired page and feeds the page to the fulltext-index.
The servlet accepts a parameter defining the action to perform (see defaults/web.xml, index, crawl, crawllinks)
The option check-boxes are placed in ConfigPortal.html
stream the original to the search result thumb viewer. This has two
reasons:
- animated gifs cause 100% cpu and deadlocks in the jvm gif parser; a
known bug which is obviously not yet fixed
- animated gifs now appear in the search result also as animation
so web.xml in defaults dir is applied first and optional DATA/SETTINGS/web.xml loaded on top.
By using this Jetty feature (default web.xml) we assure that changes to the default are applied to existing installations
and individual addition/changes are still respected.
fixes http://mantis.tokeek.de/view.php?id=517
The url given in bug report uses a gzip input stream which causes the HTTPClient.writeto() throw an IOException due to incomplete input stream. This in turn prevents the 302 reponse to the client browser.
By limiting to serve target content just on httpstatus=200 will proxy the header reponse and client browsers redirect settings can be honored.
preferred over https. While this is a bad idea from the standpoint of
security it is more common applicable for environments where http and
https mix and for some domains https is not available. Then the
double-check is possible even if no postprocessing is performed.
parsing into individual pages and add them all using different URLs.
These constructed urls are generated from the source url with an
appended page=<pagenumber> attribute to the url get/post properties.
This will distinguish the different page entries. The search result list
will then replace the post parameter with a url anchor # mark which
causes that the original url is presented in the search result. These
URLs can be opened directly on the correct page using pdf.js which is
now built-in into firefox. That means: if you find a search hit on page
5 and click on the search result, firefox will open the pdf viewer and
shows page 5.
from documents which actually are inside the index. This can be
reproduced using the crawl result table at
http://localhost:8090/CrawlResults.html?process=5
The cache is temporary disabled to remove the bad behaviour, however a
later reactivation of that feater may be possible.
generated dynamically which will stay static in the future. This applies
mainly to the search result favicon in front of search hits. These icons
will now be generated once, but then caches in the browser. There is
also a YaCy-internal cache for these icons which had prevented the
re-generation of the icons in YaCy, but this cache is now superfluous
since the browser should not call the servlet ViewImage again.
reger