improvements in etext results UX

[eroux]

Let's improve the UX of search results in the following way:

• [x] it's really necessary to be able to expand the context directly in the search result, to understand what context the match is in (we've had a lot of user feedback saying it's important)
• [x] there are often multiple matches in one etext, the query returns the top 3 matches per etext, it would be useful to have a way to "display more matches", or maybe having 3 excerpts when there are 3 matches but that makes the search results very busy
• [x] we need a button "open in etext viewer" (or something like that) so that when you see an excerpt that you like you can just open the etext viewer at the corresponding passage, it's also a fundamental part of the experience of searching in the etexts

[roopeux]

@berger-n , the three ideas of Élie can be covered like this:

• Initially, show only one snippet for etext results (less than now)
• Add "More" button
• When More is clicked, show up to five snippets which are twice longer than the ordinary.
• When user clicks a snippet (in either More or Less states), open the etext reader directly, pointing to the text that was clicked.

[berger-n]

ok @roopeux thanks! then we'll need to add a few things to the results, for each hit inside a given text:

• the offset of the chunk, to be able to link to somewhere like ...?startChar=3919...
• a longer snippet (note that current users get a context which is far longer than 2x the ordinary, see link then Expand context)

I'm also wondering about the use cases of the etext search, isn't it supposed to provide exhaustive results in a given text? which could be available from a page like https://library-dev.bdrc.io/show/bdr:MW1PD152291 (displaying a Find in this text field but no results by default)

[eroux]

thanks @berger-n !

Normally the results contain the whole matching chunks (just not in the highlights section unfortunately), perhaps that would be enough for the full snippet? The full chunks are objects that also contain the startChar / endChar

Currently the only search implemented is the general search, but yes we need to implement the full search in an etext or etext instance.

In theory restricting the search to a certain etext or etext instance should be as simple as using associatedRes

[berger-n]

oh ok I see, thanks! so I'm afraid something's missing, here's the first result for https://library-dev.bdrc.io/osearch/search?q=spyod%20%27jug&etext_search%5B0%5D=true:

{
    "_id": "MW00EGS1016761",
    "_index": "bdrc_prod",
    "_score": 0.023055173,
    "_source": {
        "altLabel_bo_x_ewts": [
            "dpal gsang chen rnying ma'i chos spyod zab bcud rin po che'i gter mdzod/"
        ],
        "altLabel_en": [
            "Dpal gsang chen rnying ma'i chos spyod zab bcud rin po che'i gter mdzod =A collection of daily practices of Nyingmapa school /by Dpal-ldan-shes-rab."
        ],
        "associated_res": [
            "WA00EGS1016761",
            "T1082",
            "T225",
            "PR3JW7543"
        ],
        "authorshipStatement_en": [
            "by mkhan-chen dpal-ldan-shes-rab"
        ],
        "comment_en": [
            "collection of devotional works of the rnying ma tradition",
            "tbrc copy missing second title page",
            "Scanned at M/S Satluj Infotech Images, E-45, Sector 27 Noida, District Gautam Buddha Nagar, U.P. 201301 via New Delhi, India for the Buddhist Digital Resource Center."
        ],
        "creation_date": "2006-12-04T18:39:54.187Z",
        "db_score": 0.5002899,
        "db_score_in_type": 0.5002899,
        "etext_access": 3,
        "etext_quality": 2,
        "firstScanSyncDate": "2016-03-30T16:20:30.571Z",
        "graphs": [
            "MW00EGS1016761"
        ],
        "join_field": "instance",
        "language": [
            "LangBo"
        ],
        "merged": [
            "WA00EGS1016761",
            "W00EGS1016761"
        ],
        "other_id": [
            "014253880",
            "BQ7662.6 .D65 2000",
            "2001293793",
            "008982196-3"
        ],
        "pop_score": 0.4045376181602478,
        "pop_score_in_type": 0.4045376181602478,
        "pop_score_rk": 0.4045376181602478,
        "prefLabel_bo_x_ewts": [
            "dpal gsang chen rnying ma sgo chen rgyal ba'i ri khrod bdud 'dul gsang sngags gling dgon gyi chos spyod gces btus zab bcud rin po che'i gter mdzod/"
        ],
        "prefLabel_en": [
            "the treasury of jewels : a collection of essential nyingmapa prayers and practices"
        ],
        "printMethod": [
            "PrintMethod_Modern"
        ],
        "publicationDate": "2000",
        "publisherLocation_en": [
            "sarnath, varanasi"
        ],
        "publisherName_en": [
            "pema samye chokor ling"
        ],
        "scans_access": 5,
        "scans_freshness": 0.023588656,
        "script": [
            "ScriptTibt"
        ],
        "summary_en": [
            "Collection of Nyingma prayers and practices, edited by Gochen Khenpo Palden Sherab (1938-2010)."
        ],
        "type": [
            "Instance"
        ],
        "workGenre": [
            "T225"
        ],
        "workIsAbout": [
            "T1082"
        ]
    },
    "inner_hits": {
        "etext": {
            "hits": {
                "hits": [
                    {
                        "_id": "UTIE0OPI40212F00_I01JW66",
                        "_index": "bdrc_prod",
                        "_routing": "MW00EGS1016761",
                        "_score": 2.3948097,
                        "_source": {},
                        "highlight": {
                            "chunks.text_bo": [
                                "བསྟོད་དང། གནས་བརྟན\nབཅུ་དྲུག་ཕྱག་མཆོད། 51-63 འཕགས་པ་བཀྲ་ཤིས་བརྒྱད་པ། 63-66 ཕྱོགས་བཅུའི་མུན་སེལ། 66-68 <em>སྤྱོད</em>",
                                "་<em>འཇུག</em>་མཆོད་པདང། མཎྜལ་སོ་བདུན་མ། 69-74 སྡོམ་གསུམ་བླ་\n\n༄༅།། བརྒྱུད་གསོལ་འདེབས། 74-88 ཨོ་རྒྱན་བླ་མ་རྒྱངས",
                                "<em>སྤྱོད</em>་<em>འཇུག</em>་ལས་གསུངས་པའི་མཆོད་པ་འབུལ་བའི་ཚུལ་ནི། རིན་ཆེན་སེམས་དེ་བཟུང་བར་བྱ་བའི་ཕྱིར། དེ་བཞིན་གཤེགས་པ་",
                                "བགྲོད་པའི་གྲུ་བོ་ཆེ་དར་གཅིག་ཏུ་གནས་པར་བྱ་བའི་བློས་བཟའ་བར་བྱ་ཡི། \nཆགས་པ་དང་སྲེད་པ་འཕེལ་བའི་བློས་མི་བཟའོ། <em>སྤྱོད</em>",
                                "་<em>འཇུག</em>་ལས། རན་པ་ཙམ་དུ་བཟའ་བར་བྱ། ཞེས་པས་དེའི་ཚུལ་ནི་ཡན་ལག་བརྒྱད་པར་ལུས་ཀྱི་ཆ་གཉིས་ཟས། གཅིག་སྐོམ། གཅིག་"
                            ]
                        }
                    }
                ],
                "max_score": 2.3948097,
                "total": {
                    "relation": "eq",
                    "value": 1
                }
            }
        }
    }
}

[eroux]

ah, that's interesting, something's missing indeed... it should be fixed in the new version I just deployed

[berger-n]

not far from done: https://library-dev.bdrc.io/osearch/search?q=spyod%20%27jug&etext_search%5B0%5D=true

---

note that when the snippet is expanded, the highlighting is made in the client and thus not perfect:

---

it wouldn't hurt to have a See all results button next to More, would it? though I couldn't make the associated_res query work for now (link)

[roopeux]

This looks good to me. Is there something still to be done?

[berger-n]

fixed highlighting in expanded matches: https://library-dev.bdrc.io/osearch/search?q=tshal%20bu%20AND%20dbu%20med&associatedTradition%5B0%5D=TraditionNyingma