Acces rate limitations to this search mode by unauthenticated users are
set low by default to prevent unwanted server overload but can be
customized through the SearchAccessRate_p.html configuration page
Fixes#291
Previously search navigators/facets elements were sorted only by counts.
Now from the ConfigSearchPage_p.html admin page, sort direction
(ascending/descending) and type (on counts or labels) can be customized
independently for each navigator.
On some conditions (especially when reaching timeout), concurrent Solr
query tasks used by the /HostBrowser.html and /api/linkstructure.json
never terminated, thus leaking resources, as reported by @Vort in issue
#246
New "Media Type detection" section in the advanced crawl start page
allow to choose between :
- not loading URLs with unknown or unsupported file extension without
checking the actual Media Type (relying Content-Type header for now).
This was the old default behavior, faster, but not really accurate.
- always cross check URL file extension against the actual Media Type.
This lets properly parse URLs ending with an apparently odd file
extension, but which have actually a supported Media Type such as
text/html.
Sample URLs with misleading file extensions added as documentation in
the crawl start page.
fixes issue #244
- Removed calls to no more existing clearResources functions (on PDFont
class and its children) since upgrade to pdfbox 2.n.n
- Removed hacky usage of protected internal ClassLoader function. This
removes the warnings displayed when running with JDK9 or JDK10 :
[java] WARNING: Illegal reflective access by
net.yacy.document.parser.pdfParser$ResourceCleaner (file:<path>) to
method java.lang.ClassLoader.findLoadedClass(java.lang.String)
[java] WARNING: Please consider reporting this to the maintainers
of net.yacy.document.parser.pdfParser$ResourceCleaner
[java] WARNING: Use --illegal-access=warn to enable warnings of
further illegal reflective access operations
[java] WARNING: All illegal access operations will be denied in a
future release
Crawling thousands of pdf documents from various sources after
modifications applied, revealed no new memory leak related to pdfbox
(measurements done with JVisualVM).
Was useless as done in an already synchronized block, and the lock
object was assigned a new value in that same block, and nowhere else a
lock is requested on that same object.
When using the 'From Link-List of URL' as a crawl start, with lists in
the order of one or more thousands of links, the failreason_s Solr field
maximum size (32kb) was exceeded by the string representation of the URL
must-match filter when a crawl URL was rejected because not matching.
SimpleDateFormat must not be used by concurrent threads without
synchronization for parsing or formating dates as it is not thread-safe
(internally holds a calendar instance that is not synchronized).
Prefer now DateTimeFormatter when possible as it is thread-safe without
concurrent access performance bottleneck (does not internally use
synchronization locks).
This makes possbile to set up much more advanced document crawl filters,
by filtering on one or more document indexed fields before inserting in
the index.
For a better control on the maximum simultaneous outgoing http
connections, as already done for any other http connections (crawls, rwi
search, p2p protocol) using the net.yacy.cora.protocol.http.HTTPClient
If not interested in displaying this on your search results and notably
on a peer with limited resources this can help saving some CPU and
outgoing network connections.
Initializing Thread names using the Thread constructor parameter is
faster as it already sets a thread name even if no customized one is
given, while an additional call to the Thread.setName() function
internally do synchronized access, eventually runs access check on the
security manager and performs a native call.
Profiling a running YaCy server revealed that the total processing time
spent on Thread.setName() for a typical p2p search was in the range of
seconds.
That was caused by concurrent modifications (with addHighlightField()
function) to the same SolrQuery instance when requesting Solr on remote
peers in p2p search.
By not generating MD5 hashes on all words of indexed texts, processing
time is reduced by 30 to 50% on indexed documents with more than 1Mbytes
of plain text.
if at least one of the image size fields is enabled in index (images_height_val,
images_width_val, images_pixel_val).
Previously all fields were required to be enabled (hint: default setting
is height + width enabled)
Otherwise once this operation is applied, the remote Solr(s) instances
are deconnected and the embedded Solr is connected even if disabled by
setting "core.service.fulltext".
Also use constants for related default setting values.
- Use the EnhancedXMLResponseWriter only when requested output is "exml"
- Use the Standard Solr writers when possible, for example for json, xml
or javabin output formats
- Return an error when the requested format can not been rendered with
an external Solr server only
Important : this modification is necessary for peers using exclusively
an external Solr server to be reachable as robinson targets in p2p
search, as the binary format ("javabin") is the default Solr exchange
format for peers.
Before this, when a peer requested a remote one attached only to an
external Solr (no embedded one), it ended with "Invalid type" error, as
the remote peer answered with xml although binary format was requested.
This is necessary when you want to attach to a dedicated external Solr
server protected with basic http authentication and requested over https
but having only a self-signed certificate.
The current default boost function
(`recip(ms(NOW,last_modified),3.16e-11,1,1)`) for the Date ranking
profile is indeed working fine.
What can trigger the error `unexpected docvalues type NUMERIC for field
'last_modified'` is the previous default boost function (quite old now)
or any custom one using the Solr `ord` or `rord` functions on the
last_modified field.
Then the problem was that the migration code in the Switchboard supposed
to detect the old date boost function was incorrect (one trailing right
parenthesis in excess), so the deprecated function remained.
This fixes issue #169.
Previously, when clicking a selected facet in the search results page to
unselect it, all other eventually selected modifiers/facets were also
removed.
With the appropriate vocabulary settings in Vocabulary_p.html page, this
can produce Vocabulary search facets displaying item types referenced in
html documents by microdata annotation.
Tested notably, but not limited to, vocabulary classes/types defined by
Schema.org and Dublin Core.
check case insensitive.
As keywords are compared lower case, make sure user input keyword:Key
or keyword:key will be shown as active in facet entry key.
Thus allowing to choose at configuration or per search request, whether
extending or not results beyond strict content domain filter (image,
video, audio or application).
Related graphical controls to be added to user interface.
Required for proper operation when the default system locale is Turkish,
as dottless and dotted i characters have specific case conversion rules
in this language.
Introduced through the new configurable setting
network.unit.protocol.https.preferred, defaulting to false for now.
Let choose to prefer using https when available on remote peers to
perform YaCy protocol operations including notably hello or transferRWI.
Not yet implemented for every YaCy protocol operations.
When a crawl is started, a new field to exclude content from scraping is
available. The field can be identified with the class name of div tags.
All text contained in such a div tag where the configured class name(s)
match are not indexed, while the remaining page is indexed.
Allow typing directly internationalized domain names including non ASCII
characters in the search field.
Search is done using the ASCII Compatible Encoding (ACE) representation.
Default is still http to prevent any regressions, but a new setting is
available to choose https as the preferred protocol to perform remote
searches.
New configuration setting 'remotesearch.https.preferred' is manually
editable in yacy.conf file or in Advanced Properties page
(/ConfigProperties_p.html).
Should be enabled as default in the future for improved privacy.
Https could also eventually be used for other peers communications.
Also ensure authentication is not lost by Digest timeout when navigating
between index.html and search results page.
This way, running searches with extended features on a remote peer or a
password protected peer works with a regular user (with "Extended
search" rights).
When authenticating on the search page with a user without "Extended
search" rights, it appears as authenticated, but has just its usual
access to the public search features.
Thus allowing a more convenient way (wihout the need to go to the admin
section) to login when searching on your remote or password protected
peer and benefit from extended search features such as Heuristics,
Bookmarking or JavasScript resorting.
Can be disabled using the ConfigSearchPage_p.html.
The SearchEvent listen to changes on each of its navigators, and the
information about their overall state is sent with each fetched search
item (as a "data-nav-generation" attribute). Then the browser can
regularly fetch a fresh version of yacysearchtrailer.html only if
necessary (when that nav-generation value change).
When the limit is reached, a button allow expanding/collapsing remaining
tags.
When this feature is activated without a limit to the number of
displayed tags, when encountering search results with a very large
number of keywords, the results page can become almost unusable (very
long vertical scrollbar)
This is a fix for mantis 766 ( http://mantis.tokeek.de/view.php?id=766 )
Since the upgrade to Digest authentication, access to protected search
features was indeed disabled once the Digest nonce timed out.
After Digest auth timeout the browser no more sent authentication
information and as the search results page is not private, protected
features were simply be hidden without asking browser again for
authentication.
Adding a supplementary parameter when accessing the search results as
authenticated fixes this.
Inspired from the existing one used on image search, and consistent with
post filtering on content domain applied in SearchEvent.addNodes().
These filters are quite simplistic but at least audio, video or
application search now return results. Previously, when filtering on
these content domains, many results pages (and often even the first
page) were empty while the total results count suggested that results
should be available. This was because filtering on domain was only
applied AFTER requesting Solr indexes.
- added some missing increments from RWI results
- decrement relevant navigator counts when solr or RWI results are
evicted because duplicates detection or constraints checked belatedly
- do not compute facets when unnecessary to avoid unwanted CPU load
- do not increment from facets when already done
- do not rely on facets on remote solr peers requests, as most of the
time only a limited part of their total results if fetched (thus also
preventing unnecessary load on remote peers)
- use a concurrency friendly score map for the dates navigators to
prevent unwanted ConcurrentModificationExceptions
This improves the situation for the most obvious inconsistencies in
search navigators counts, but more has to be done for a true accuracy
(notably when query modifiers constraints are applied belatedly - after
the solr or RWI retrieval request - such as the content domain
constraint)
Was inadequately modified in my previous related commits (making next
pages buttons unavailable in Search portal mode), as
SearchEvent.local_solr_available did not count the total filtered
results but only the ones within the currently fetched result page(s).
This modification has indeed low incidence as eventual query modifiers
are already applied when requesting the local solr index.
It mainly impact doublons detected with results from remote peers.
Also updated javadocs for clarification.
As a server-side oriented alternative to the JavaScript realtime
resorting feature proposed in PR #104.
The goal is the same as in this PR : having the possibility compensate
the network latency of various peers results fetching and obtain once
possible a consistently ranked result set.
to make sure updated documents are indexed with their last-modified
date as provided in current crawl.
(to patch moddate always with firstseen might bear the risk of miss
actual updates).
As reported by davide on YaCy forums (
http://forum.yacy-websuche.de/viewtopic.php?f=23&t=6004 ) when the
system is on high load, unless reading carefully YaCy configuration
file, it could be difficult to understand why remote search results are
not fetched.
Especially for Turkish speaking users using "tr" as their system default
locale : strings for technical stuff (URLs, tag names, constants...)
must not be lower cased with the default locale, as 'I' doesn't becomes
'i' like in other locales such as "en", but becomes 'ı'.
Using a Reentrant lock instead of the intrinsic synchronization lock
permits limiting the blocking time to acquire a lock.
Useful on a very busy Cache concurrently accessed by many threads : when
the time to acquire a lock is too high, getting/storing content on the
cache becomes inefficient, and it is then better to fall back to loading
remote resources.
Illustrated by the CacheTest stress test and some traces reported in
mantis 751 ( http://mantis.tokeek.de/view.php?id=751 )
Also add when possible a warning level log message on input stream
closing error instead of failing silently. This could help understanding
some IO exceptions such as "too many files open".
This enables keyword navigator to filter on keywords. Added search page
output and layout config for keywords, allowing e.g. in Intranet use
to display the keywords. No styling or links applied to the keyword
text (but is desirable possibly in combination with bootstrap-tagsinput
for future/intranet).
When Webgraph Solr core is enabled, crawling and removing from index an
URL whose hash starts with the '-' character (example URL :
https://cs.wikipedia.org/ whose hash is "-2-HuTEndn4x") produced a full
ParseException stack trace in YaCy logs. This was not blocking because
the Solr query parser is able to escape itself the query and run it
successfully, but filled uselessly YaCy logs.
Added as an additional icon with title in the search progress bar, to
inform about background search feeder threads terminated or still
running. While giving a bit more information to users about the p2p
search process, this can help choosing whether or not wait a little bit
more time before going to the next page, in order to get results from
various sources sorted as best as possible (see #91 for a discussion
about sorting accuracy and network latency).
Other related modifications included :
- regular updates to statistics in the progress bar until the
background feeders are completely terminated.
- removed some uses of unsecure and discouraged JavaScript elements
luccioman