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.
Ensuring http post method is used for operations with server-side
effects (in respect of http semantics), and a valid transaction token is
provided by the user-agent.
- to prevent unwanted exposure of index entries about private
local/intranet documents when switching from "Intranet Indexing" mode
while attached to remote Solr instance(s)
- to warn user about remote Solr instance(s) still attached when
switching from modes other than "Intranet Indexing"
- properly handle IPv6 loopback address replacement
- replace loopback address or host only when accessing peer remotely
- replace loopback part with the peer hostname as requested rather than
with its seed public IP as this works better for Intranet mode and when
peer is behind a reverse proxy.
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.
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.
When inlined (for example in the CrawlProfileEditor_p.html page) :
search only on the comment, as the url is not visible
On regular display : search on comment OR url, instead of comment AND
url. Otherwise searching on comments terms is almost useless as these
terms are not necessarily present in the url.
Applied the same default results page size as when a type filter is
defined for proper and consistend page navigation when combining type
filter and search query.
Otherwise, when interested in viewing `Link List` for example, each time
you typed a new URL, `Parsed Sentences` view mode was selected as
default and you had to selected again the view mode you are insterested
in.
As discussed in issue #160 , blacklist entries can indeed currently not
be "complete" regular expressions, but must be structured as a domain
part, a separator character ('/'), and a path part.
- Fixes issue #160 : handle properly syntax exceptions with a user
friendly message
- Fixes loss of information on multiple blacklist entries editions
- Fixes loss of entries when moving entries from one list to another
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.
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.
Upgraded to InetAccessHandler.
Added InetPathAccessHandler extension to InetAccessHandler to maintain
path patterns capability previously available in IPAccessHandler but
lost in InetAccessHandler.
Filtering on IPv6 addresses is now supported.
Support for deprecated pattern formats such as "192.168." and
"192.168.1.1/path" has been removed, but startup automated migration
should convert such patterns eventually present in serverClient.
Previously named with their ISO 3166-1 country code : this way, when
setting language to "Browser" in ConfigBasic.html, it didn't work
properly when browser preferred language was Chinese or Greek as their
respective language codes are "zh" and "el" (not "cn" and "gr" which are
their country codes)
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.
Otherwise, when authenticated as admin and navigating from search
results or admin pages to the search start page (/index.html), if
nothing is done on that page within HTTP Digest Auth timeout (about
2mn), then search is performed without authentication and so without
extended search features.
Thus eventually including the same optional login link/status in the
search start page than in the results page, for the same convenient
login without the need to use the Administration section.
Restores the behavior introduced eleven years ago (see commit
479861a3cf) and lost by mistake 3 years
ago (see commit 617dd9c97b), when the
click handler started referencing a missing HTML id.
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.
Resizing JPEG snapshot images through /api/snapshot.jpg failed when
running on OpenJDK, but rendered successfully with a Oracle JDK.
Details in mantis 772 ( http://mantis.tokeek.de/view.php?id=772 ).
Removing any alpha component (useless in snapshot images) from the
rendered resized image solves the issue.
Previously rendered as a broken URL containing the absolute file path of
a snapshot on the search server.
Now rendered as a valid URL linking to the /api/snapshot API to provide
available snapshot content. Snapshot format is selected among the
available ones in the following order of preference : JPG/PNG, PDF, and
XML.
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.
If swichted on link (click) to the tag adds the keyword to the search query.
If a keyword navigator is active the selected keyword adds or replaces
a query keyword: modifier (currently replace was choosen as multiple
keywords are not fully supported yet)
- 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).
Using unfiltered detailed counts (local and remote entries found before
doubles detection and before applying query modifiers) was confusing and
inconsistent with the total count. It could let think more results are
to come in the next pages, without understanding why they are not
displayed.
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.
Previously, when checking for the first time the robots.txt policy on a
unknown host (not cached in the robots table), result was always empty
in the /getpageinfo_p.xml api and in the /CrawlCheck_p.html page. Next
calls returned however the correct information.
Complements the recent modification related to images in commit 7f395ef.
Unfortunately many documents metadata fetched from the freeworld p2p
network have only partial information about embedded images. Without
proper error handling, this made many searches in p2p mode to fail
completely.
This should be a help to make a preview of search results.
The image is computed from the list of embedded images, it is
always the first image in that list.
In rss-type results the image is presented like
<media:content medium="image" url="https://abc.xyz/logo.png"/>
as defined in
http://www.rssboard.org/media-rss#media-content
This allow large files parsing and preview, while preventing unwanted
OutOfMemory errors which are likely to occur when adding to the Solr
Index resources larger than configured crawler limits.
Thus enable getpageinfo_p API to return something in a reasonable amount
of time on resources over MegaBytes size range.
Support added first with the generic XML parser, for other formats
regular crawler limits apply as usual.
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.
Otherwise on a malformed getpageinfo_p XML response (from the browser
point of view), JavaScript errors where thrown and the ajax status
steering wheel remained displayed indefinitely.
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'.
This prevent rendering a big and inconvenient scrollbar on resources
containing many links.
If really needed, preview of all links is still available with a "Show
all links" button.
Doesn't affect the number of links used once the crawl is effectively
started, as the list is then loaded again server-side.
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).
Redirections set for the transition of any eventual external uses:
- /api/getpageinfo.xml to /api/getpageinfo_p.xml
- /api/getpageinfo.json to /api/getpageinfo_p.json
Replaced by shortcuts defined by the HTML "accesskey" attribute which
has the advantage to be advertised by screen readers when focusing the
corresponding buttons, contrary to custom JavasScript key handlers.
Now With Firefox :
- "Alt + Shift + n" for next page
- "Alt + Shift + p" for previous page
Following ARIA recommendation : "keyboard shortcuts enhance, not
replace, standard keyboard access." ( see
https://www.w3.org/TR/wai-aria-practices/#kbd_shortcuts_behavior_design)
Fix for mantis 711 (http://mantis.tokeek.de/view.php?id=711)
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
- added the new setting as configurable in the "Debug/Analysis" settings
page. Debug/analysis is its main purpose for now as there is currently
no nice and "understansable" ranking score info servlet (see forum
discussion http://forum.yacy-websuche.de/viewtopic.php?f=8&t=5884 )
- render in the "Search Page Layout" page preview when enabled
- added constants
When import thread is terminated :
- now stop refreshing and stay on the monitoring page to give user a
feedback after a long running import
- added link to the next monitoring step : results from surrogates
reader
- added link to new import
On the new import page, added a link on the eventual last import report.
Take into account the already existing default limit value (especially
useful after a long crawl or surrogates import), or a custom one from
parameter "count".
Added a "Show all" link for convenience.
When using a public HTTP URL in /IndexImportMediawiki_p.html, the remote
file now is directly streamed and processed, allowing import of several
GB dumps even with a low memory remote peer, and without need to
manually download the dump file first.
Using an HTML "file" input was confusing (as reported by promocore on
YaCy forum : http://forum.yacy-websuche.de/viewtopic.php?f=5&t=5965) ,
and it only worked with MS IE/Edge on a local YaCy peer :
- for security reasons some current major browsers such as Firefox or
Chrome do not allow to send full file path information when using a file
form input
- the local file system selection popup doesn't make sense when you
want to import a dump on a remote YaCy server
count.
This might be tangential related to http://mantis.tokeek.de/view.php?id=736
as the example includes a local index search, while rwi results are not
counted.
- ensure use of HTTP POST method when performing server side effect
operations
- transaction token required to ensure the request has effectively been
requested by user interaction
- enabled HTTP POST calls with Digest HTTP authentication
- made API calls compatible with API newly restricted to HTTP POST only
with transaction token validation
- ensured backward compatibility with older entries recorded as HTTP
GET
- ensure use of HTTP POST method : HTTP GET should only be used for
information retrieval and not to perform server side effect operations
(see HTTP standard https://tools.ietf.org/html/rfc7231#section-4.2.1)
- a transaction token is now required for these administrative form
submissions to ensure the request can not be included in an external
site and performed silently/by mistake by the user browser
A port value of -1 will disable this option.
If set to a value greater 0, YaCy listens on this of on the local loopback
address (127.0.0.1) for a shutdown or restart signal.
E.g. connect to http://localhost:8005/shutdown will stop the YaCy server.
http://localhost:8005/restart will restart it.
This option allows to stop YaCy locally independant from the web web
frontend (which might be configured for password protected remote access).
This is a fix for mantis 715 (http://mantis.tokeek.de/view.php?id=715).
A possible path scenario that could leading to this case :
- YaCy is running low in memory
- a search is requested
- before the end of search results rendering, the cleanup job runs and
deletes the running search event from the cache because of short memory
- then yacysearchitem renders with "-UNRESOLVED_PATTERN-" parameter
values passed to the statistics() JavaScript function
HTTP "Referer" header sent by the browser when using YaCy can now be
controlled either with the referrer meta tag as a global policy, or only
for search result links by adding the attribute rel="noreferrer".
To improve privacy with the less possible regressions, the default is
set as meta tag with value "origin-when-cross-origin" : internal YaCy
links behavior is not affected, but when visiting external websites
referrer url is not empty but stripped from query parameters and path.
Older browsers, Safari, MS IE and Edge do not support the referrer meta
tag, so the standard but less flexible noreferrer link type can also be
enabled as an alternative.
User-friendly settings page to be implemented.
- Added a new method to check activation of mandatory fields on
Collection Configuration commit, consistently with checks previously
performed in Switchboard startup and with mandatory fields in the
default schema.
- Reorganized default schema and CollectionConfiguration enumeration :
moved no more mandatory fields in a specific section, and moved fields
enabled at startup to the mandatory section.
- Marked mandatory fields as required and with stronger font in the
IndexSchema_p.html page
As noticed by @reger24, abusive use of OpenSearch systems should be
prevented, especially if allowing to parse and reuse HTML results.
robots.txt file is now checked before requesting an external OpenSearch
system to respect the host exclusions and eventual crawl-delay value.
The check is also performed when trying to add a new OpenSearch URL
template through the /ConfigHeuristics_p.html admin page.
As discussed in PR #93 with @JeremyRand and @reger24 this new advanced
settings page includes:
- a new setting to control remote Solr responses encoding
- some existing debug settings which could not be set through the admin
user interface
This property has been deprecated four years ago by commit
d74472f562. For any active search event
id, it was then always filled with "-UNRESOLVED_PATTERN-".
This is to be more easily understandable and to reflect more accurately
the current memory strategies implementations that eventually set the
"proper" state not only because DHT reception.
As described in mantis 725 (http://mantis.tokeek.de/view.php?id=725) the
HostBrowser.xml api directory entries had incorrect count attribute
value.
This was because the HostBrowser html page and backing template servlet
evolved, but modifications were not reported on the xml api.
This will prevent mistakenly hiding a div element not designed to be an
infobox but having a ".info" parent (After having previously added the
possibility for a div - and not only a span element - to be an infobox).
As mentioned in issue #103, control settings over YaCy disk usage
already existed but lacked a user-friendly way to set them.
I added it to the Performance_p.html administration page with a little
refactoring on the "Resource Observer" fieldset for improved
accessibility and HTML standards respect.
Also added the possibility to enable/disable the autoregulation fonction
from this page.
Made use of ol, li, thead, th, tbody, h1 and h2 html tags.
Added aria-label attributes to provide alternative textual information
previously only conveyed by color cue.
Tested behavior with NVDA 2016.4 screen reader.
luccioman
