(stackexchange-converter) Create tool for converting stackexchange 7z-files to digestible sqlite db:s

code/features-convert/stackexchange-xml/readme.md

@@ -0,0 +1,19 @@
+Stackexchange's data is a jumble of questions and answers,
+where the answers refer to the questions with a parentId field.
+
+e.g.
+```xml
+<?xml version="1.0" encoding="utf-8"?>
+<posts>
+  <row Id="1" PostTypeId="1" AcceptedAnswerId="51" CreationDate="2016-01-12T18:45:19.963" Score="10" ViewCount="424" Body="&lt;p&gt;When I've printed an object I've had to choose between high resolution and quick prints.  What techniques or technologies can I use or deploy to speed up my high resolution prints?&lt;/p&gt;&#xA;" OwnerUserId="16" LastActivityDate="2017-10-31T02:31:08.560" Title="How to obtain high resolution prints in a shorter period of time?" Tags="&lt;resolution&gt;&lt;speed&gt;&lt;quality&gt;" AnswerCount="2" CommentCount="6" ContentLicense="CC BY-SA 3.0" />
+  <row Id="2" PostTypeId="1" AcceptedAnswerId="12" CreationDate="2016-01-12T18:45:51.287" Score="34" ViewCount="7377" Body="&lt;p&gt;I would like to buy a 3D printer, but I'm concerned about the health risks that are associated with its operation. Some groups of scientists say it can be &lt;a href=&quot;http://www.techworld.com/news/personal-tech/scientists-warn-of-3d-printing-health-effects-as-tech-hits-high-street-3460992/&quot;&gt;harmful&lt;/a&gt; for humans.&lt;/p&gt;&#xA;&#xA;&lt;p&gt;What do I need to consider before buying a 3D printer if I care about my health? Are there any safe printers?&lt;/p&gt;&#xA;" OwnerUserId="20" LastEditorUserId="334" LastEditDate="2016-11-15T16:16:11.163" LastActivityDate="2019-06-10T23:18:34.190" Title="Is 3D printing safe for your health?" Tags="&lt;print-material&gt;&lt;safety&gt;&lt;health&gt;" AnswerCount="4" CommentCount="1" ContentLicense="CC BY-SA 3.0" />
+  <row Id="12" PostTypeId="2" ParentId="2" CreationDate="2016-01-12T19:13:00.710" Score="23" Body="&lt;p&gt;There is very little information about safety available, as home 3D printers are relatively new. However, plastics such as ABS have a long history in making plastic products, and a study found..." />
+</posts>
+```
+
+Since the search engine wants to extract keywords for each thread
+holistically, not by question or answer, it is necessary to re-arrange
+the data (which is very large).  SQLite does a decent job of enabling
+this task.
+
+See [tools/stackexchange-converter](../../tools/stackexchange-converter).

code/features-convert/stackexchange-xml/src/main/java/nu/marginalia/integration/stackexchange/sqlite/StackExchangePostsDb.java

@@ -33,8 +33,9 @@ public class StackExchangePostsDb {
 
     /** Construct a SQLIte file containing the Posts in the stack exchange-style 7z file */
     @SneakyThrows
-    public static void create(Path sqliteFile,
-                       Path stackExchange7zFile) {
+    public static void create(String domain,
+                              Path sqliteFile,
+                              Path stackExchange7zFile) {
         if (Files.exists(sqliteFile))
             Files.delete(sqliteFile);
         String connStr = "jdbc:sqlite:" + sqliteFile;
@@ -58,6 +59,13 @@ public class StackExchangePostsDb {
                     stackExchange7zFile
             );
 
+            var insertMeta = connection.prepareStatement("""
+                    INSERT INTO metadata(domainName)
+                    VALUES (?)
+                    """);
+            insertMeta.setString(1, domain);
+            insertMeta.executeUpdate();
+
             var insertPost = connection.prepareStatement("""
                      INSERT INTO post(id, threadId, postYear, title, body, origSize, tags)
                      VALUES (?, ?, ?, ?, ?, ?, ?)

code/features-convert/stackexchange-xml/src/main/resources/db/stackexchange.sql

@@ -8,4 +8,8 @@ CREATE TABLE post (
     tags TEXT
 );
 
+CREATE TABLE metadata (
+   domainName TEXT
+);
+
 CREATE INDEX post_threadId ON post(threadId);

code/tools/stackexchange-converter/build.gradle

@@ -0,0 +1,52 @@
+plugins {
+    id 'java'
+    id "io.freefair.lombok" version "8.2.2"
+    id 'application'
+
+    id 'jvm-test-suite'
+}
+
+java {
+    toolchain {
+        languageVersion.set(JavaLanguageVersion.of(20))
+    }
+}
+
+application {
+    mainClass = 'nu.marginalia.tools.StackexchangeConverter'
+    applicationName = 'stackexchange-converter'
+}
+
+tasks.distZip.enabled = false
+
+dependencies {
+    implementation project(':code:features-convert:stackexchange-xml')
+
+    implementation libs.lombok
+    annotationProcessor libs.lombok
+    implementation libs.bundles.slf4j
+    implementation libs.notnull
+
+    implementation libs.guice
+    implementation libs.jsoup
+    implementation libs.trove
+    implementation libs.fastutil
+
+    implementation libs.bundles.nlp
+    implementation libs.commons.lang3
+
+    testImplementation libs.bundles.slf4j.test
+    testImplementation libs.bundles.junit
+    testImplementation libs.mockito
+}
+
+
+test {
+    useJUnitPlatform()
+}
+
+task fastTests(type: Test) {
+    useJUnitPlatform {
+        excludeTags "slow"
+    }
+}

code/tools/stackexchange-converter/readme.md

@@ -0,0 +1,24 @@
+This tool converts from stackexchange's 7z-compressed XML
+format to a sqlite database that is digestible by the search engine.
+
+See [features-convert/stackexchange-xml](../../features-convert/stackexchange-xml) for
+an explanation why this is necessary. 
+
+Stackexchange's data dumps can be downloaded from archive.org
+here: [https://archive.org/details/stackexchange](https://archive.org/details/stackexchange)
+
+<b>Usage</b>
+
+```shell
+$ stackexchange-converter domain-name input.7z output.db
+```
+
+Stackexchange is relatively conservative about allowing
+new questions, so this is a job that doesn't run more than once.
+
+<b>Note</b>:  Reading and writing these db files is *absurdly* slow
+on a mechanical hard-drive.
+
+## See Also
+
+* [features-convert/stackexchange-xml](../../features-convert/stackexchange-xml)

code/tools/stackexchange-converter/src/main/java/nu/marginalia/tools/StackexchangeConverter.java

@@ -0,0 +1,31 @@
+package nu.marginalia.tools;
+
+import nu.marginalia.integration.stackexchange.sqlite.StackExchangePostsDb;
+
+import java.nio.file.Files;
+import java.nio.file.Path;
+
+public class StackexchangeConverter {
+    public static void main(String[] args) {
+
+        if (args.length != 3) {
+            System.err.println("Converts a stackexchange Posts 7z file to a Marginalia-digestible sqlite-db\n");
+            System.err.println("Arguments: domain-name input-file.7z output-file.db");
+            return;
+        }
+
+        String domain = args[0];
+
+        Path inputFile = Path.of(args[1]);
+        Path outputFile = Path.of(args[2]);
+
+        if (!Files.exists(inputFile))
+            System.err.println("Input file " + inputFile + " does not exists");
+
+        System.out.println("Converting " + inputFile);
+
+        StackExchangePostsDb.create(domain, outputFile, inputFile);
+
+        System.out.println("... done!");
+    }
+}

settings.gradle

@@ -75,6 +75,7 @@ include 'code:tools:experiment-runner'
 include 'code:tools:website-adjacencies-calculator'
 include 'code:tools:screenshot-capture-tool'
 include 'code:tools:load-test'
+include 'code:tools:stackexchange-converter'
 
 include 'third-party:porterstemmer'
 include 'third-party:xz'