[tor-commits] [metrics-web/master] Imports UpdateNews.java

[karsten at torproject.org]

commit 664195d24da6408707aecda856fb7f6f9ff3c494
Author: Iain R. Learmonth <irl at fsfe.org>
Date:   Tue Mar 20 14:11:49 2018 +0000

    Imports UpdateNews.java
    
    This commit imports UpdateNews.java verbatim, as attached to #23854 by
    Karsten.
    
    https://trac.torproject.org/projects/tor/attachment/ticket/23854/UpdateNews.java
---
 .../org/torproject/metrics/web/UpdateNews.java     | 131 +++++++++++++++++++++
 1 file changed, 131 insertions(+)

diff --git a/src/main/java/org/torproject/metrics/web/UpdateNews.java b/src/main/java/org/torproject/metrics/web/UpdateNews.java
new file mode 100644
index 0000000..527988f
--- /dev/null
+++ b/src/main/java/org/torproject/metrics/web/UpdateNews.java
@@ -0,0 +1,131 @@
+package org.torproject.metrics.web;
+
+import com.google.gson.Gson;
+import com.google.gson.GsonBuilder;
+
+import java.io.BufferedReader;
+import java.io.FileWriter;
+import java.io.InputStreamReader;
+import java.net.URL;
+import java.util.ArrayList;
+import java.util.List;
+
+public class UpdateNews {
+  private static class News {
+    String start;
+    String end;
+    Boolean ongoing;
+    List<String> places;
+    List<String> protocols;
+    String description;
+    List<String> links;
+    Boolean unknown;
+  }
+
+  /** Update news. */
+  public static void main(String[] args) throws Exception {
+    URL textFile = new URL(
+        "https://trac.torproject.org/projects/tor/wiki/doc/"
+        + "MetricsTimeline?format=txt");
+    Gson gson = new GsonBuilder().disableHtmlEscaping().create();
+    List<News> news = new ArrayList<>();
+    try (BufferedReader br = new BufferedReader(new InputStreamReader(
+        textFile.openStream()))) {
+      String line;
+      Boolean unknown = null;
+      while ((line = br.readLine()) != null) {
+        if (line.startsWith("== Unknown")) {
+          unknown = true;
+          continue;
+        }
+        if (!line.startsWith("||") || line.startsWith("||=start")) {
+          continue;
+        }
+        line = line.trim()
+            .replaceAll("×", "×")
+            .replaceAll("§", "§")
+            .replaceAll("–", "–")
+            .replaceAll("“", "“")
+            .replaceAll("”", "”");
+        String[] parts = line.split("\\|\\|");
+        News entry = new News();
+        entry.start = parts[1].replaceAll("~", "").trim();
+        if (entry.start.contains(" ")) {
+          entry.start = entry.start.substring(0, entry.start.indexOf(" "));
+        } else if (entry.start.isEmpty()) {
+          entry.start = null;
+        }
+        entry.end = parts[2].replaceAll("~", "").trim();
+        if ("ongoing".equalsIgnoreCase(entry.end)) {
+          entry.end = null;
+          entry.ongoing = true;
+        } else if (entry.end.contains(" ")) {
+          entry.end = entry.end.substring(0, entry.end.indexOf(" "));
+        } else if (entry.end.isEmpty()) {
+          entry.end = null;
+        }
+        for (String place : parts[3].trim().split(" +")) {
+          if (!place.isEmpty()) {
+            if (null == entry.places) {
+              entry.places = new ArrayList<>();
+            }
+            entry.places.add(place);
+          }
+        }
+        for (String protocol : parts[4].trim().split(" +")) {
+          if (!protocol.isEmpty()) {
+            if (null == entry.protocols) {
+              entry.protocols = new ArrayList<>();
+            }
+            entry.protocols.add(protocol);
+          }
+        }
+        String desc = parts[5].trim();
+        while (desc.contains("[")) {
+          int open = desc.indexOf("[");
+          int space = desc.indexOf(" ", open);
+          int close = desc.indexOf("]", open);
+          if (open < 0 || space < 0 || close < 0) {
+            System.err.println("Cannot convert link.");
+            System.exit(1);
+          }
+          desc = desc.substring(0, open) + "<a href=\""
+              + desc.substring(open + 1, space) + "\">"
+              + desc.substring(space + 1, close) + "</a>"
+              + desc.substring(close + 1);
+        }
+        while (desc.contains("`")) {
+          int open = desc.indexOf("`");
+          int close = desc.indexOf("`", open + 1);
+          if (open < 0 || close < 0) {
+            System.err.println("Cannot convert code fragment.");
+            System.exit(1);
+          }
+          desc = desc.substring(0, open) + "<code>"
+              + desc.substring(open + 1, close) + "</code>"
+              + desc.substring(close + 1);
+        }
+        entry.description = desc;
+        if (parts.length >= 7) {
+          for (String link : parts[6].split("[\\[\\]]")) {
+            link = link.trim();
+            if (link.isEmpty()) {
+              continue;
+            }
+            if (null == entry.links) {
+              entry.links = new ArrayList<>();
+            }
+            entry.links.add("<a href=\"" + link.substring(0, link.indexOf(" "))
+                + "\">" + link.substring(link.indexOf(" ") + 1) + "</a>");
+          }
+        }
+        entry.unknown = unknown;
+        news.add(entry);
+      }
+    }
+    try (FileWriter fw = new FileWriter(
+        "src/main/resources/web/json/news-raw.json")) {
+      fw.write(gson.toJson(news));
+    }
+  }
+}