How to trigger specific Google Custom Search Engine refinement label? - google-custom-search

Currently, our organization is using Google Custom Search engine to provide auto suggestion, and we have about 3 refinement labels configured in our CSE. Previously, we're using WebSearch and SearchControl, and the WebSearch has a setSiteRestriction method which allows us to specifically select a refinement label:
- http://code.google.com/apis/websearch/docs/reference.html#_class_GwebSearch
Previous code example:
var searchControl = new google.search.SearchControl();
var webSearch = new google.search.WebSearch();
//Refinement allows us to tell Google which specific sites to search
var refinement="Support";
//filter custom search and currently there are 3 refinements
(some other variables declaration here including 'product')
switch(product)
{
case "10000":
refinement = "Support1";
break;
case "10200":
refinement = "Support1";
break;
case "10001":
refinement = "Support2";
break;
default:
break;
}
/*this is the code to fill in the custom search. The refinement was set above - either "Support", "Support1", or "Support2".*/
webSearch.setSiteRestriction('cseId', refinement);
......
However, currently we're migrating to CustomSearchControl tool to replace the deprecated WebSearch, but apparently I couldn't find any way to specifically select a refinement label based on the value of switch case statement. Immediate help needed here, and if there's a relevant docs that you guys can point me to will be much appreciated. thanks! :)

When using customSearchControl, you can set the refinement label in the options.
This will
a) not restrict other refinements in your search to the keywords you
may have added with ('more: ' + refinement), and
b) also highlight
the refinement tab to tell the user what you did on their behalf.
var customSearchOptions =
{ 'defaultToRefinement' : 'refinement_label_name' };
var customSearchControl =
new google.search.CustomSearchControl('YOUR_CSE_ID', customSearchOptions);
The defaultToRefinement parameter is mentioned in the Custom Search Element JavaScript API Reference.
Previously, this answer was introduced here.

Got the answer. Appended the following lines to the code:
var customSearchControl = new google.search.CustomSearchControl(cseId);
customSearchControl.setSearchStartingCallback(this, function(control, searcher, query)
{
searcher.setQueryAddition('more:' + refinement);
});

Related

Google Docs script, return text value in the middle of specified text

[current result][1]
[1]: https://i.stack.imgur.com/O8AGa.png
The problem is that they all apply to the same line
I want to get the values individually from the same line
Rule
▶random text◁
→ Change only values between ▶ and ◁
Script currently in use
var body = DocumentApp.getActiveDocument().getBody();
var foundElement = body.findText('▶([^\S]+)◁');
while (foundElement != null) {
// Get the text object from the element
var foundText = foundElement.getElement().asText();
// Where in the element is the found text?
var start = foundElement.getStartOffset();
var end = foundElement.getEndOffsetInclusive();
// Set Bold
foundText.setBold(start, end, true);
// Change the Foreground color
foundText.setForegroundColor(start, end, "#ff326d");
// Find the next match
foundElement = body.findText('▶([^\S]+)◁', foundElement);
}
Some regular expressions are not fully supported
As per the documentation A subset of the JavaScript regular expression features are not fully supported, such as capture groups and mode modifiers.
I'd recommend to open a feature request on Google Issue Tracker as recommended here

Programmatically update the signal for a multi-click in vega/vega-lite

Following the example on the website: https://vega.github.io/editor/#/examples/vega-lite/interactive_bar_select_highlight
I want to programmatically set the selections via signals. I realize that I could emulate a click by doing the following
VEGA_DEBUG.view.signal("select_tuple", {"unit":"","fields":[{"type":"E","field":"_vgsid_"}],"values":[1]})
However, I cannot proceed to select another, e.g., the shift select of the 2
VEGA_DEBUG.view.signal("select_tuple", {"unit":"","fields":[{"type":"E","field":"_vgsid_"}],"values":[2]})
This makes sense, since only shift-click accumulates the state.
I tried modifying the accumulated signal
VEGA_DEBUG.view.signal("select", {"_vgsid_":[1,2],"vlMulti":{"or":[{"_vgsid_":1},{"_vgsid_":2}]}})
However, this does not help. Is this not possible? I understand that a custom solution may be possible in hand-rolled vega, as opposed to that compiled from vega-lite.
Thanks.
Just need to set VEGA_DEBUG.view.signal("select_toggle", true) before adding the new select!!
After much research I made this example of how to change the vega-lite brush programmatically
https://observablehq.com/#john-guerra/update-vega-lite-brush-programmatically
Using #koaning example this stack overflow question I figured that you can change the brush by updating "brush_y" (assuming that your selection is called brush) or change the selection using "brush_tuple" (which doesn't seem to update the brush mark)
viewof chart = {
const brush = vl.selectInterval("brush").encodings("y");
const base = vl
.markBar()
.select(brush)
.encode(
vl.x().count(),
vl.y().fieldQ("Horsepower"),
vl.color().if(brush, vl.value("steelblue")).value("gray")
)
.height(maxY);
return base.data(data).render();
}
update = {
// From https://codepen.io/keckelt/pen/bGNQPYq?editors=1111
// brush_y -> brush_tuple -> brush
// Updates on pixels
chart.signal("brush_y", [by0, maxY / 2]);
await chart.runAsync();
}
Crossposting here in case it might be useful for anyone

Remove "MIME type" column from Filent Content List

I am Using a Script Adapter by passing payload to get contend for a Content List from "Search with values" event
When Contend get loaded to content list , i have a custom view to preview them. But if i clicked on MIME type column , It opens a separate view with the mapped viewer
So I need to remove this column or make it un-clickable
1) I am passing search values to content list's "Search with values" event , from where can i handle Content List's contend loading ,any Dojo Event i can use ?
2) With Script Adapter can i do this without going for a "response filter"
Edit :
As Nicely explained by "Ivo Jonker" (in his answer - "or try to specifically locate the widgets on your page" and with his example code)
responsed = page.ContentList8.ecmContentList.getResultSet();
var cols = responsed.structure.cells[0];
for (i=cols.length-1; i>0; i--){
var col = cols[i];
if (col.field=="mimeTypeIcon")
cols.splice(i,1);
}
page.ContentList78.ecmContentList.setResultSet(responsed);
I simply remove this row. Thanks Again and lovely blog , hope you keep posting more great articles.
The values passed through the Search With Values event will eventually be handled by the icm.pgwidget.contentlist.dijit.DocumentSearchHandler
that in turn creates a SearchTemplate to execute the search (ecm.model.SearchTemplate.prototype.search). One option would be to aspect/before/around the DocumentSearchHandler#query to manipulat the searchresults and by that way to remove the column.
The wiring however does not provide any handles to achieve this for a specific query-resultset combination leaving you to either fix this on a global scale (icm.pgwidget.contentlist.dijit.DocumentSearchHandler.prototype#query), or try to specifically locate the widgets on your page.
Personally, taking into account #2, i'd go for the responsefilter-option if you feel the global solution wouldn't be a problem, or alternatively i'd personally prefer to create a simple ICM widget that instantiates/implements a "plain" ecm.widget.listView.ContentList and exposes a wire to set the ecm.model.Resultset.
You'd then be able to create your own Searchquery in a scriptadapter, remove the column, and pass the resultset.
The script adapter could be something like:
var scriptadapter=this;
var queryParams={};
queryParams.query = "SELECT * FROM Document where id in /*your list*/";
queryParams.retrieveAllVersions = false;
queryParams.retrieveLatestVersion = true;
queryParams.repository = ecm.model.desktop.repositories[0];
queryParams.resultsDisplay = {
"sortBy": "{NAME}",
"sortAsc": true,
"columns": ["{NAME}"],
"honorNameProperty": true};
var searchQuery = new ecm.model.SearchQuery(queryParams);
searchQuery.search(function(response/*ecm.model.Resultset*/){
//remove the mimeTypeIcon
var cols = response.structure.cells[0];
for (i=cols.length-1; i>0; i--){
var col = cols[i];
if (col.field=="mimeTypeIcon")
cols.splice(i,1);
}
//emit the resultset to your new contentlist, be sure to block the regular synchrounous output of the scriptadapter
scriptadapter.onPublishEvent("icm.SendEventPayload",response);
//The contentlist wire would simply do contentlist.setResultSet(response);
});

How to clean existing properties and replace with metadata template on Photoshop (scripting)?

While creating a script that would automate all the different tasks I do when I start working on a new picture on Photoshop, I encountered the following problem.
Manually, I would Ctrl + Alt + Shift + I, click on the template I want and choose the option "Clear existing properties and replace with template properties".
I can't find the way to do precisely this. The best thing I managed to find is something like this :
app.activeDocument.info.author = "test";
app.activeDocument.info.caption = "";
app.activeDocument.info.captionWriter = "";
app.activeDocument.info.headline = "";
app.activeDocument.info.instructions = "";
app.activeDocument.info.keywords = "";
app.activeDocument.info.authorPosition = "";
app.activeDocument.info.credit = "";
app.activeDocument.info.source = "";
app.activeDocument.info.category = "";
app.activeDocument.info.supplementalCategories = "";
app.activeDocument.info.title = "";
// etc.
And it actually doesn't really work like the "Clear existing properties and replace with template properties".
I didn't find anything on the Photoshop scripting guide, nor on the internet. Any help would be greatly appreciated !
What I think is the problem is Photoshop separates file-metadata from its activeDocument-metadata. What you see in "File info..." (via Ctrl+Alt+Shift+I) is supposed to represent the file in the filesystem, which metadata is embedded in.
There are several scripting guides to Photoshop scripting. I think the one relevant for you would be "Javascript Tools Guide", specifically the chapter 10 "Scripting Access to XMP Metadata".
Is it important for you to set up the metadata already when creating a new picture? If not, you may want to look at a solution using a customized export script.
It customizes XMP-metadata upon exporting like
Create a basic metadata object:
var meta = new XMPMeta();
Provide a namespaceURI (see XMP specs) known to photoshop along with tag name, and value:
meta.setProperty(XMPConst.NS_XMP, "CreatorTool", app.version);
Save the image temporarily (using other script):
var imgFile = new File(fileName);
saveImage(fileName);
Finish saving by adding the metadata-object:
var metaFile = new XMPFile(imgFile.fsName, XMPConst.FILE_UNKNOWN, XMPConst.OPEN_FOR_UPDATE);
if (metaFile.canPutXMP(meta)) { metaFile.putXMP(meta); }
metaFile.closeFile(XMPConst.CLOSE_UPDATE_SAFELY);
Doing it this way also erases any existing or default metadata.

drupal multiple node forms on one page

I have a view that displays several nodes. I want to place node form below each displayed node. Both node_add and drupal_get_form directly in template.php works fine, but I get forms with same form ID of NODETYPE_node_form and validation and submitting does not work as expected.
If you had to put several node forms on one page, what would be your general approach?
Progress so far...
in template.php while preprocessing node
$author_profile and $content is set before.
$unique = $vars['node']->nid;
$node = new StdClass();
$node->uid = $vars['user']->uid;
$node->name = $vars['user']->name;
$node->type = 'review';
$node->language = '';
$node->title = t('Review of ') . $vars['node']->realname . t(' by ') . $vars['user']->realname . t(' on ') . $content->title;
$node->field_review_to_A[0]['nid'] = $nodeA->nid;
$node->field_review_to_B[0]['nid'] = $vars['node']->nid;
$node->field_review_to_profile[0]['nid'] = $author_profile->nid;
if(!function_exists("node_object_prepare")) {
include_once(drupal_get_path('module', 'node') . '/node.pages.inc');
}
//$vars['A_review_form'] = drupal_get_form('review_node_form', $node);
$vars['A_review_form'] = mymodule_view($node, $unique);
in mymodule module
function mymodule_view($node, $unique) {
if(!function_exists("node_object_prepare")) {
include_once(drupal_get_path('module', 'node') . '/node.pages.inc');
}
$output = drupal_get_form('review_node_form_' . $unique, $node);
return $output;
}
function mymodule_forms($form_id, $args) {
$forms = array();
if (strpos($form_id, "review_node_form_") === 0) {
$forms[$form_id] = array('callback' => 'node_form');
}
return $forms;
}
function mymodule_form_alter(&$form, $form_state, $form_id) {
if (isset($form['type']) && isset($form['#node']) && $form_id != $form['type']['#value'] .'_node_form' && $form['type']['#value'] == 'review') {
$type = content_types($form['#node']->type);
if (!empty($type['fields'])) {
module_load_include('inc', 'content', 'includes/content.node_form');
$form = array_merge($form, content_form($form, $form_state));
}
$form['#pre_render'][] = 'content_alter_extra_weights';
$form['#content_extra_fields'] = $type['extra'];
//$form['#id'] = $form_id;
//$form['#validate'][0] = $form_id . '_validate';
$form['title']['#type'] = 'value';
$form['field_review_to_A']['#type'] = 'value';
$form['field_review_to_B']['#type'] = 'value';
$form['field_review_to_profile']['#type'] = 'value';
}
}
Questions
My take on summarizing unclear questions...
Is this good general approach for displaying multiple node forms on same page?
Is it OK to copy/paste code from content modules content_form_alter function in function mymodule_form_alter? Will it not brake things if content module is updated?
Should i set $form['#id']? Without it all forms has same HTML form ID of node_form, with it ID is unique, like review_node_form_254. Thing is that there is no difference of how form is submitted. Setting $form['#validate'][0] does not seem to influence things too. Maybe I should set $form[button]['#submit'][0] to something else? Now its node_form_submit.
Why validation does not work even with unique form id and form validate function? If i submit last form without required field all previous forms gets particular fields red. should I make my own validation function? what level - form or field? Any tips on where to start?
You need to implement hook_forms() to map different ids to the same builder function.
The NODETYPE_node_form ids you mention are already an example of this mechanism, as they are all mapped to the same builder function (node_form()) within the node modules node_forms() implementation.
You can find links to more examples in the 'Parameters' explanation off the drupal_get_form() function.
This was exceptionally useful to me.
Documentation on all the drupal APIs is so lacking - I was tearing my hair out. The crucial bit for me, that I don't think is covered anywhere else on the net:
CCK adds its fields to your form through hook_form_alter() . But it does this based on the form_id. So if the form ID is different (cause you want multiple ones on the same page), you need to replicate a bit of the CCK code to add the fields to your form regardless.
That is what this does brilliantly.
I really did not dig to the bottom of it, but it seems to me that you pretty much did all the relevant digging by yourself.
From what I understand by inspecting the code only, you are right in thinking that content_form_alter() is the show-stopper.
Maybe this is a naïve suggestion, but what about writing your own implementation of hook_form_alter() starting from the content_form_alter() code but changing the conditions that make the alteration to occur? I am thinking to something along these lines:
function modulename_form_alter(&$form, $form_state, $form_id) {
if (isset($form['type']) && isset($form['#node']) &&
stripos($form_id, $form['type']['#value'] .'_node_form')) {
...
}
}
I did not actually test the code above, but I hope it is self-explenatory: you actually trigger the changes for a give customised type of content if MYCCKTYPE_node_form is a substring of the actual form_id (es: MYCCKTYPE_node_form_234).
Hope this helps at least a bit... Good luck! :)
EDIT: TWO MORE THINGS
It just occurred to me that since your custom implementation of hook_form_alter() will live aside the CCK one, you would like to add a check to avoid the form to be modified twice something like:
&& $form_id != $form['type']['#value'] .'_node_form'
The other way you might try to implement this by not using CCK but implementing you custom type programmatically (this might even be an advantage if you plan to use your module on various different sites).