From 8cb51b79225c32232a136e8becf12016eb9dfc16 Mon Sep 17 00:00:00 2001
From: vinogradovnet <dmitry@vinogradov.net>
Date: Sat, 8 Aug 2026 02:54:20 +0300
Subject: [PATCH] feat(compression): add Russian language pack (#9581)

Merge-train validated (tip 6ce4effef8). Vitest failures confirmed as base-red (#9679).
---
 .../services/compression/languageDetector.ts  |   1 +
 .../compression/rules/ru/context.json         |  38 +++++++
 .../services/compression/rules/ru/dedup.json  |  30 ++++++
 .../services/compression/rules/ru/filler.json |  86 +++++++++++++++
 .../compression/rules/ru/structural.json      | 101 ++++++++++++++++++
 .../services/compression/rules/ru/ultra.json  |  46 ++++++++
 6 files changed, 302 insertions(+)
 create mode 100644 open-sse/services/compression/rules/ru/context.json
 create mode 100644 open-sse/services/compression/rules/ru/dedup.json
 create mode 100644 open-sse/services/compression/rules/ru/filler.json
 create mode 100644 open-sse/services/compression/rules/ru/structural.json
 create mode 100644 open-sse/services/compression/rules/ru/ultra.json

diff --git a/open-sse/services/compression/languageDetector.ts b/open-sse/services/compression/languageDetector.ts
index d44295d225..9e1851443c 100644
--- a/open-sse/services/compression/languageDetector.ts
+++ b/open-sse/services/compression/languageDetector.ts
@@ -7,6 +7,7 @@ const LANGUAGE_HINTS: Record<string, RegExp[]> = {
   es: [/\b(?:necesito|archivo|codigo|código|fallo|gracias|puedes)\b/i],
   de: [/\b(?:ich|datei|fehler|bitte|kannst|konfiguration|danke)\b/i],
   fr: [/\b(?:fichier|erreur|merci|peux|besoin)\b/i],
+  ru: [/\b(?:\u044d\u0442\u043e|\u0447\u0442\u043e|\u043a\u0430\u043a|\u0435\u0441\u043b\u0438|\u0447\u0442\u043e\u0431\u044b|\u043a\u043e\u0442\u043e\u0440\u044b\u0439|\u043c\u043e\u0436\u0435\u0442|\u043d\u0443\u0436\u043d\u043e|\u0435\u0441\u0442\u044c|\u0431\u044b\u043b\u043e|\u0431\u0443\u0434\u0435\u0442|\u043c\u043e\u0436\u043d\u043e|\u0434\u043e\u043b\u0436\u0435\u043d|\u0444\u0430\u0439\u043b|\u043e\u0448\u0438\u0431\u043a\u0430|\u043f\u0440\u043e\u0431\u043b\u0435\u043c\u0430|\u0434\u0430\u043d\u043d\u044b\u0435)\b/i, /[\u0430-\u044f\u0451]/i],
   ja: [/[\u3040-\u30ff]/],
   id: [/\b(?:saya|kamu|anda|dengan|untuk|yang|tidak|bisa|terima\s+kasih|dari)\b/i],
 };
diff --git a/open-sse/services/compression/rules/ru/context.json b/open-sse/services/compression/rules/ru/context.json
new file mode 100644
index 0000000000..26534688ff
--- /dev/null
+++ b/open-sse/services/compression/rules/ru/context.json
@@ -0,0 +1,38 @@
+{
+  "language": "ru",
+  "category": "context",
+  "rules": [
+    {
+      "name": "subject_omission",
+      "pattern": "^(?:Я |Мы |Вы )(?:можем|должны|будем|хотим|нужно)\\b\\s*",
+      "replacement": "",
+      "context": "all",
+      "category": "context",
+      "minIntensity": "full"
+    },
+    {
+      "name": "known_fact_hedging",
+      "pattern": "(?<=\\.)\\s*(?:Возможно|Наверное|Может быть),\\s+",
+      "replacement": "",
+      "context": "assistant",
+      "category": "context",
+      "minIntensity": "full"
+    },
+    {
+      "name": "redundant_clarification",
+      "pattern": "\\b(?:как я уже говорил|как уже упоминалось|как было сказано)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "context",
+      "minIntensity": "full"
+    },
+    {
+      "name": "obvious_continuation",
+      "pattern": "\\b(?:далее|затем|после этого|в итоге)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "context",
+      "minIntensity": "ultra"
+    }
+  ]
+}
diff --git a/open-sse/services/compression/rules/ru/dedup.json b/open-sse/services/compression/rules/ru/dedup.json
new file mode 100644
index 0000000000..6678979858
--- /dev/null
+++ b/open-sse/services/compression/rules/ru/dedup.json
@@ -0,0 +1,30 @@
+{
+  "language": "ru",
+  "category": "dedup",
+  "rules": [
+    {
+      "name": "thought_repetition",
+      "pattern": "([^.!?]+[.!?])\\s+\\1",
+      "replacement": "$1",
+      "context": "all",
+      "category": "dedup",
+      "minIntensity": "full"
+    },
+    {
+      "name": "word_duplication",
+      "pattern": "\\b(\\w+)\\s+\\1\\b",
+      "replacement": "$1",
+      "context": "all",
+      "category": "dedup",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "synonymous_repetition",
+      "pattern": "\\b(проблема|ошибка)\\b[^.!?]*\\b(проблема|ошибка)\\b",
+      "replacement": "$1",
+      "context": "all",
+      "category": "dedup",
+      "minIntensity": "full"
+    }
+  ]
+}
diff --git a/open-sse/services/compression/rules/ru/filler.json b/open-sse/services/compression/rules/ru/filler.json
new file mode 100644
index 0000000000..aa28ae6153
--- /dev/null
+++ b/open-sse/services/compression/rules/ru/filler.json
@@ -0,0 +1,86 @@
+{
+  "language": "ru",
+  "category": "filler",
+  "rules": [
+    {
+      "name": "pleasantries",
+      "pattern": "\\b(?:конечно|с радостью|рад помочь|могу помочь|обязательно|безусловно|разумеется)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "polite_framing",
+      "pattern": "\\b(?:пожалуйста|если хотите|если можно|будьте добры|будьте любезны|прошу вас)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "verbal_wrapping",
+      "pattern": "\\b(?:давайте разберём|давайте посмотрим|попробуем разобраться|постараюсь помочь)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "assistant",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "hedging",
+      "pattern": "\\b(?:возможно|наверное|может быть|скорее всего|вероятно|видимо|похоже)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "filler_adverbs",
+      "pattern": "\\b(?:в целом|на самом деле|в принципе|как правило|по сути|фактически|буквально)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "empty_qualifiers",
+      "pattern": "\\b(?:удобный|хороший|эффективный|мощный|отличный|прекрасный)(?!\\s+(?:вариант|способ|решение|метод|инструмент))\\b",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "redundant_openers",
+      "pattern": "^(?:Привет|Здравствуйте|Добрый день|Доброе утро|Добрый вечер)\\s*[,.!?\\s]?\\s*",
+      "replacement": "",
+      "context": "user",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "excessive_gratitude",
+      "pattern": "\\b(?:Большое спасибо|Огромное спасибо|Спасибо заранее|Заранее благодарю|Очень признателен)\\b[,.!?\\s]*",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "softeners",
+      "pattern": "\\b(?:немного|немножко|чуть-чуть|слегка|несколько|как-то)\\b\\s*",
+      "replacement": "",
+      "context": "all",
+      "category": "filler",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "assistant_fillers",
+      "pattern": "^(?:Вот|Ниже|Это|Здесь)\\s+(?:есть|находится)?\\s*",
+      "replacement": "",
+      "context": "assistant",
+      "category": "filler",
+      "minIntensity": "lite"
+    }
+  ]
+}
diff --git a/open-sse/services/compression/rules/ru/structural.json b/open-sse/services/compression/rules/ru/structural.json
new file mode 100644
index 0000000000..78bf745f29
--- /dev/null
+++ b/open-sse/services/compression/rules/ru/structural.json
@@ -0,0 +1,101 @@
+{
+  "language": "ru",
+  "category": "structural",
+  "rules": [
+    {
+      "name": "problem_phrasing",
+      "pattern": "\\b(?:проблема заключается в том, что|дело в том, что|суть в том, что)\\b\\s*",
+      "replacement": "проблема: ",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "full"
+    },
+    {
+      "name": "causality_verbose",
+      "pattern": "\\b(?:это приводит к тому, что|это означает, что|из этого следует, что)\\b\\s*",
+      "replacement": "→ ",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "full"
+    },
+    {
+      "name": "purpose_phrases",
+      "pattern": "\\b(?:для того чтобы|с целью того чтобы)\\b\\s*",
+      "replacement": "чтобы ",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "causality_phrases",
+      "pattern": "\\b(?:в связи с тем, что|по причине того, что|ввиду того, что)\\b\\s*",
+      "replacement": "из-за ",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "full"
+    },
+    {
+      "name": "concession_phrases",
+      "pattern": "\\b(?:несмотря на то, что|хотя и)\\b\\s*",
+      "replacement": "хотя ",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "note_phrases",
+      "pattern": "\\b(?:стоит отметить, что|следует иметь в виду, что|важно понимать, что|необходимо учитывать, что)\\b\\s*",
+      "replacement": "",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "full"
+    },
+    {
+      "name": "redundant_directive",
+      "pattern": "\\b(?:важно помнить|не забывайте|помните о том)\\b\\s*",
+      "replacement": "",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "full"
+    },
+    {
+      "name": "approximation",
+      "pattern": "\\b(?:примерно|приблизительно)\\b\\s*",
+      "replacement": "≈ ",
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "full"
+    },
+    {
+      "name": "forbidden_abbreviations_dots",
+      "pattern": "\\b(?:т\\.к\\.|т\\.е\\.|и т\\.д\\.|и т\\.п\\.|см\\.|напр\\.|и др\\.|в т\\.ч\\.)\\b",
+      "replacement": "",
+      "replacementMap": {
+        "т.к.": "так как",
+        "т.е.": "то есть",
+        "и т.д.": "",
+        "и т.п.": "",
+        "см.": "см",
+        "напр.": "например",
+        "и др.": "",
+        "в т.ч.": ""
+      },
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "lite"
+    },
+    {
+      "name": "forbidden_abbreviations_dash",
+      "pattern": "\\b(?:кол-во|к-рый|св-во)\\b",
+      "replacement": "",
+      "replacementMap": {
+        "кол-во": "количество",
+        "к-рый": "который",
+        "св-во": "свойство"
+      },
+      "context": "all",
+      "category": "structural",
+      "minIntensity": "lite"
+    }
+  ]
+}
diff --git a/open-sse/services/compression/rules/ru/ultra.json b/open-sse/services/compression/rules/ru/ultra.json
new file mode 100644
index 0000000000..ca2609aac7
--- /dev/null
+++ b/open-sse/services/compression/rules/ru/ultra.json
@@ -0,0 +1,46 @@
+{
+  "language": "ru",
+  "category": "ultra",
+  "rules": [
+    {
+      "name": "ultra_compression_conjunctions",
+      "pattern": "\\b(?:однако|тем не менее|в то время как)\\b\\s*",
+      "replacement": "—",
+      "context": "all",
+      "category": "ultra",
+      "minIntensity": "ultra"
+    },
+    {
+      "name": "ultra_compression_articles",
+      "pattern": "\\b(?:является|представляет собой)\\b\\s*",
+      "replacement": "—",
+      "context": "all",
+      "category": "ultra",
+      "minIntensity": "ultra"
+    },
+    {
+      "name": "ultra_compression_verbs",
+      "pattern": "\\b(?:необходимо|требуется|нужно)\\b\\s*",
+      "replacement": "",
+      "context": "all",
+      "category": "ultra",
+      "minIntensity": "ultra"
+    },
+    {
+      "name": "ultra_punctuation",
+      "pattern": "[,:;]\\s+",
+      "replacement": " ",
+      "context": "all",
+      "category": "ultra",
+      "minIntensity": "notes"
+    },
+    {
+      "name": "ultra_lowercase",
+      "pattern": "(?<=\\.)\\s+([А-ЯЁ])",
+      "replacement": " $1",
+      "context": "all",
+      "category": "ultra",
+      "minIntensity": "notes"
+    }
+  ]
+}