- BaseX-Talk - mailman.uni-konstanz.de

XML Parsing Error: syntax error, Location: http://localhost:8984/rest/oshb?query=count(.), Line Number 1, Column 1:
by Patrick Durusau 17 Mar '22

17 Mar '22

Greetings! What puzzles is the correct count is returned, despite the error message, with the cursor under the first digit. Looking in the response, it appears that "favicon.ico" is missing and I suspect is the cause of the error. But I wanted to confirm that assumption. Thanks! Patrick -- Patrick Durusau patrick(a)durusau.net Technical Advisory Board, OASIS (TAB) Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300 Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps) Another Word For It (blog): http://tm.durusau.net Homepage: http://www.durusau.net Twitter: patrickDurusau

2 2

Dumb question: Adding files to BaseX database
by Patrick Durusau 14 Mar '22

14 Mar '22

Greetings! I'm using db:add to add files to a BaseX database, but am I just missing how to do that in the GUI? Thanks! Patrick -- Patrick Durusau patrick(a)durusau.net Technical Advisory Board, OASIS (TAB) Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300 Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps) Another Word For It (blog): http://tm.durusau.net Homepage: http://www.durusau.net Twitter: patrickDurusau

3 3

Way to Dynamically import a module?
by Buddy Kresge 14 Mar '22

14 Mar '22

I was wondering if in BaseX there is a way at run time to import a .xqm module, outside of the traditional "import module" statement. Thanks in advance for any guidance you can provide. Buddy

2 3

Fuzzy matching in ft:search for terms with spaces
by Tamara Marnell 10 Mar '22

10 Mar '22

Hi everyone, The results for ft:search() with the "fuzzy" option set to "true" are unexpected for phrases with spaces in them. Usually "fuzzy" set to "false" will return fewer search results than when it's set to true, but I'm finding the opposite for phrases. Sometimes ft:search() doesn't return known exact matches for a phrase unless "fuzzy" is "false." For example, in my development environment I have a document titled Test Finding Aid. I'm tokenizing all text() nodes in an index database and verified that "test finding aid" is a phrase in the node included for full-text. <tokens> test xml test finding aid 1964 2002 test finding aid finding aid prepared central oregon community college 2008 [etc.] </tokens> When I ft:search() for the complete phrase "test finding aid" as one term with "fuzzy" set to "true," I get 4 results, none of which are the document titled Test Finding Aid with the index entry above. They're all results with "West" in the title that happen to have "west finding aid" in them when tokenized. But when I set "fuzzy" to "false," I get 1 result, and it is for the document titled Test Finding Aid. I expect all exact results to be included in fuzzy results, but other searches show that the exact results are often more numerous than the fuzzy ones for phrases. league women voters (as 3 distinct terms) Fuzzy true: 430 results Fuzzy false: 373 results (expected) league women voters (as 1 term) Fuzzy true: 281 results Fuzzy false: 299 results (not expected) pacific coast (as 2 distinct terms) Fuzzy true: 2,551 results Fuzzy false: 1,866 results (expected) pacific coast (as 1 term) Fuzzy true: 893 results Fuzzy false: 935 results (not expected) Does anyone know what's causing this, and what can be done about it? -Tamara -- Tamara Marnell Program Manager, Systems Orbis Cascade Alliance (orbiscascade.org <https://www.orbiscascade.org/>) Pronouns: she/her/hers

2 3

How to return/use the value of a nested counter?
by Ben Engbers 10 Mar '22

10 Mar '22

Hi, I have a collection of 740 documents with the following structure: <?xml version="1.0" encoding="utf-8"?> <officiele-publicatie xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://technische-documentatie.oep.overheid.nl/schema/op-xsd-2012-1"> <metadata> <meta name="OVERHEIDop.externMetadataRecord" scheme="" content="https://zoek.officielebekendmakingen.nl/h-tk-20202021-102-2/metadata.xml" /> </metadata> <handelingen> <agendapunt> <spreekbeurt nieuw="nee"> <tekst status="goed"> <al-groep> <al>Allereerst hebben we het traditionele mondelinge vragenuur. </al> </al-groep> </tekst> </spreekbeurt> <spreekbeurt nieuw="ja"> <tekst status="goed"> <al-groep> <al>Voorzitter. Het was altijd al een eer om hier te staan.</al> </al-groep> <al-groep> <al>De vragen die ik ga stellen, gaan over stikstof.</al> </al-groep> <al-groep> <al>We zijn allemaal 100 kilometer per uur gaan rijden, maar er is nog geen gram ammoniak uit de veehouderij minder uitgestoten.</al> </al-groep> </tekst> </spreekbeurt> </spreekbeurt> <spreekbeurt nieuw="nee"> <tekst status="goed"> <al-groep> <al>U heeft helaas maar één vraag, meneer Ephraim, als Groep Van Haga.</al> </al-groep> <al-groep> <al>Ik wil de minister bedanken voor haar beantwoording.</al> </al-groep> </tekst> </spreekbeurt> </agendapunt> </handelingen> </officiele-publicatie> I am trying to concatenate all the //<al> childs from <spreekbeurt> elements. Together with an ID that I construct from //meta/@content and a counter for <spreekbeurt>, I want this output: 20202021-102-2, 1, Allereerst ... 20202021-102-2, 2, Voorzitter... + De vragen ... + We zijn ... 20202021-102-2, 3, U heeft ... + Ik wil.. I expected that the following XQuery-statemnt would do it. import module namespace functx = "http://www.functx.com"; for $Debat at $CountOuter in collection("Parliament") (: where $CountOuter <= 3:) let $debate-id := fn:analyze-string( $Debat/officiele-publicatie/metadata/meta/@content, "(\d{8}-\d*-\d*)")//fn:match/*:group[@nr="1"]/text() order by $debate-id for $Beurt at $CountInner in $Debat let $tekst := $Beurt//spreekbeurt//al/text() return($debate-id, $CountInner, $tekst) Instead it returns: 20202021-102-2, 1, Allereerst ...+ Voorzitter... + De vragen ... + We zijn ... + U heeft ... + Ik wil.. How can I use the value of $CountInner? Ben

2 3

XQuery versus XSL
by Ben Engbers 09 Mar '22

09 Mar '22

Hi, I have a collection of 740 documents with the following structure: <?xml version="1.0" encoding="utf-8"?> <officiele-publicatie xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://technische-documentatie.oep.overheid.nl/schema/op-xsd-2012-1"> <metadata> <meta name="OVERHEIDop.externMetadataRecord" scheme="" content="https://zoek.officielebekendmakingen.nl/h-tk-20202021-102-2/metadata.xml" /> </metadata> <handelingen> <agendapunt> <spreekbeurt nieuw="nee"> <tekst status="goed"> <al-groep> <al>Allereerst hebben we het traditionele mondelinge vragenuur. </al> </al-groep> </tekst> </spreekbeurt> <spreekbeurt nieuw="ja"> <tekst status="goed"> <al-groep> <al>Voorzitter. Het was altijd al een eer om hier te staan.</al> </al-groep> <al-groep> <al>De vragen die ik ga stellen, gaan over stikstof.</al> </al-groep> <al-groep> <al>We zijn allemaal 100 kilometer per uur gaan rijden, maar er is nog geen gram ammoniak uit de veehouderij minder uitgestoten.</al> </al-groep> </tekst> </spreekbeurt> </spreekbeurt> <spreekbeurt nieuw="nee"> <tekst status="goed"> <al-groep> <al>U heeft helaas maar één vraag, meneer Ephraim, als Groep Van Haga.</al> </al-groep> <al-groep> <al>Ik wil de minister bedanken voor haar beantwoording.</al> </al-groep> </tekst> </spreekbeurt> </agendapunt> </handelingen> </officiele-publicatie> I want to experiment with textmining and for these experiments, it would be usefull if for every <spreekbeurt>, all <al>/text() elements were concated.The first option is to use XQuery for concatenating. Another option is to use XSL to transform the original documents to the following structure: <?xml version="1.0" encoding="utf-8"?> <officiele-publicatie xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation="http://technische-documentatie.oep.overheid.nl/schema/op-xsd-2012-1"> <metadata> <meta name="OVERHEIDop.externMetadataRecord" scheme="" content="https://zoek.officielebekendmakingen.nl/h-tk-20202021-102-2/metadata.xml" /> </metadata> <handelingen> <agendapunt> <spreekbeurt nieuw="nee"> <tekst status="goed"> Allereerst hebben we het traditionele mondelinge vragenuur. </tekst> </spreekbeurt> <spreekbeurt nieuw="ja"> <tekst status="goed"> Voorzitter. Het was altijd al een eer om hier te staan. De vragen die ik ga stellen, gaan over stikstof. We zijn allemaal 100 kilometer per uur gaan rijden, maar er is nog geen gram ammoniak uit de veehouderij minder uitgestoten. </tekst> </spreekbeurt> </spreekbeurt> <spreekbeurt nieuw="nee"> <tekst status="goed"> U heeft helaas maar één vraag, meneer Ephraim, als Groep Van Haga. Ik wil de minister bedanken voor haar beantwoording. </tekst> </spreekbeurt> </agendapunt> </handelingen> </officiele-publicatie> Question: What are the pros and cons of both methods? Is it difficult to do this in XSL (I have only used very simple transformations)? Ben

2 1

Edit to https://docs.basex.org/wiki/Database_Server and question
by Patrick Durusau 08 Mar '22

08 Mar '22

Greetings! I edited https://docs.basex.org/wiki/Database_Server to change: Add the stop keyword to gracefully shut down the server. to: Use |basexserverstop| or |basexserverstop.bat| to gracefully shut down the server. Question, in the first paragraph following this list, the text reads: "Pressing |Ctrl+c| will close all connections and databases and gracefully shut down the server process." True? False? Somewhere in between? If true it should be the third alternative after basexserverstop.bat, but then why the scripts if we can always Ctrl+c? What am I missing? Thanks! Patrick -- Patrick Durusau patrick(a)durusau.net Technical Advisory Board, OASIS (TAB) Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300 Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps) Another Word For It (blog): http://tm.durusau.net Homepage: http://www.durusau.net Twitter: patrickDurusau

2 3

Re: [basex-talk] Keyboard issue on MacBook (Apple M1 Pro)
by Alexander Holupirek 07 Mar '22

07 Mar '22

> Hello, I'm having a small but annoying issue with keyboard input in the BaseX GUI on a new MacBook (M1 Pro). I tend to use the U.S. International keyboard layout for ease of typing accented characters: e.g., using the apostrophe/quotation key as a dead key when typing é or ü. However, with this keyboard layout, the apos/quot key doesn't register at all in the BaseX GUI. It works as expected everywhere else. I've tried switching fonts, but that doesn't seem to help. Any ideas? Thanks in advance! Tim Hi Tim, I just tried what you described on a M1. Using 'International - PC' as keyboard layout and pressing three keys, namely 'SHIFT' + 'QUOT' + 'A' gives me the following behaviour: Terminal-App (and other apps): App | Keys | Description -------------|-----------|----------------------------------------------- Terminal-App | SHIFT | (nothing happens) | " | a double quot sign (") in a (yellow) highlighted square | a | the german umlaut a (ä) on default background appears In BaseXGUI: App | Keys | Description -------------|-----------|----------------------------------------------- BaseXGUI | SHIFT | (nothing happens) | " | (nothing happens) | a | the german umlaut a (ä) on default background appears So it seems that the indication / user feedback is missing in BaseXGUI, but the umlaut character can nevertheless be entered. Do you experience the same? Alex

2 3

Results of some experiments for improving full-text search speeds
by Tamara Marnell 05 Mar '22

05 Mar '22

Good morning--I hope everyone in Europe is safe and well. This week I've been working on improving the speed of full-text searches in our website, and I wanted to share what I've found to be helpful and not so helpful. Our project: We run a website of finding aids from 47 archival repositories in the western US. Archivists submit their finding aids as EADs through a CMS that indexes each document for full-text searching in BaseX, with other custom indexes for facets like subject, material type, etc. https://archiveswest.orbiscascade.org/ Our issue: We need to perform full-text searches that match documents with all of the terms in the query. The BaseX full-text index and ft:search() function work at the text node level. So we need to use the mode "any word," then group by a unique identifier in the root and run an ft:contains() with the mode "all words," like this: for $db_id in tokenize($d, '\|') for $result in ft:search($db_id, $terms, map{'mode':'any word','fuzzy':$f}) let $ead := $result/ancestor::ead let $ark := normalize-space($ead/eadheader/eadid/@identifier) (: Other node values fetched here :) group by $ark where ft:contains(string-join($result, ' '), $terms, map{'mode':'all words','fuzzy':$f}) (: Custom ranking calculations based on other node values performed here :) return <ark>{$ark}</ark> Most users of the site are interested in their repositories only, and they use only a few words to find known collections, so most of the time results are delivered in less than 5 seconds. If users want to search for multiple terms across all 47 repositories, though, the search was taking 30-40 seconds. What I DIDN'T find helpful: Modifying the XQuery 1. Performing a separate search for each term and mapping to fetch with db:open-id() later, instead of searching for all terms at once and grouping. This was slightly slower: for $db_id in tokenize($d, '\|') let $results := fold-left( (: Get the ft:search() results as a map of node IDs and results :) ) for $node_id in map:keys($results) let $ead := db:open-id($db_id, $node_id) (: Ranking etc. :) 2. For multi-term queries, creating a temporary table to search instead of the complete full-text indexes of all databases. a. Create a database like temp123456 with a full-text index. b. For the first term, loop through all databases add the documents to temp123456 with db:add($temp_id, db:path($ead)) and optimize. c. For subsequent terms, map the results of a full-text search temp123456 instead of looping through all 47 databases, and then loop through all EADs in the temp database and remove the ones that don't match with... if (not(exists($temp_results(db:node-id($ead))))) then db:delete($temp_id, db:path($ead)) d. For the final term, return the usual ranked <ark /> values, then drop the database. This didn't work because the optimization of the temporary databases took a long time--25-30 seconds for a term with 1K document results, according to the BaseX logs. It would work in theory only if the existing full-text index could be parceled out for subsequent terms, instead of being freshly rebuilt. What I DID find helpful: Changing the server We use Amazon Web Services, so I experimented with putting the project on different instance types: General, Memory-Optimized, and Compute-Optimized. The contenders: General t2.medium: 2 vCPU, 4 GB memory General t3.large: 2 vCPU, 8 GB memory Compute-optimized c6g.xlarge: 4 vCPU, 8 GB memory Memory-optimized r6g.large: 2 vCPU, 16 GB memory At the time of testing we were on t2.medium, which is the most affordable and the smallest I thought would be sufficient. For one-term queries, the difference in speed was not that much across the instance types. For longer queries, the number of virtual CPUs and memory made a big difference. Query: native american tribes in oregon t2.medium: 55 seconds t3.large: 30 seconds r6g.large: 22 seconds c6g.xlarge: 21 seconds Memory: - Increasing from 4 GB on the t2.medium to 8 GB on the t3.large and c6g.xlarge nearly doubled the speed of my queries, even for queries of 2 words like "women's march" (20 seconds on t2.medium, 8-10 seconds on the others). - Increasing from 8 GB to 16 GB on a memory-optimized instance didn't affect the speed. The java process used only 10% of the available memory, versus 50-60% on other instance types, but the speed wasn't any faster, even increasing the maximum heap size in the basexhttp startup script. This could be because heap size isn't the right property to adjust. CPU: - Increasing from 2 vCPU to 4 vCPU on a compute-optimized instance improved speeds only for extreme queries of 4 words or more. papers t3.large: 6 seconds c6g.xlarge: 7 seconds family papers t3.large: 18 seconds c6g.xlarge: 18 seconds adams family papers t3.large: 15 seconds c6g.xlarge: 14 seconds adams family papers utah t3.large: 25 seconds c6g.xlarge: 21 seconds For most queries our users are performing, which are short and restricted to one repository, it's a negligible difference that didn't justify the increase in cost for a compute-optimized instance ($40 per month vs. $60 per month). -Tamara -- Tamara Marnell Program Manager, Systems Orbis Cascade Alliance (orbiscascade.org <https://www.orbiscascade.org/>) Pronouns: she/her/hers

3 7

Keyboard issue on MacBook (Apple M1 Pro)
by Tim Thompson 05 Mar '22

05 Mar '22

Hello, I'm having a small but annoying issue with keyboard input in the BaseX GUI on a new MacBook (M1 Pro). I tend to use the U.S. International keyboard layout for ease of typing accented characters: e.g., using the apostrophe/quotation key as a dead key when typing é or ü. However, with this keyboard layout, the apos/quot key doesn't register at all in the BaseX GUI. It works as expected everywhere else. I've tried switching fonts, but that doesn't seem to help. Any ideas? Thanks in advance! Tim -- -- Tim A. Thompson Discovery Metadata Librarian Yale University Library 344 Winchester Street New Haven, Connecticut 06511 (203) 432-4189 (office) (201) 423-9972 (mobile) www.linkedin.com/in/timathompson timothy.thompson(a)yale.edu

1 0