Fix broken build id regex & Return a set from get_filepaths (#110)

* log_analyser: Match application or title when searching for build IDs

* log_analyser: Return a set from get_filepaths

flake.lock

@@ -59,11 +59,11 @@
     },
     "nixpkgs": {
       "locked": {
-        "lastModified": 1715668745,
-        "narHash": "sha256-xp62OkRkbUDNUc6VSqH02jB0FbOS+MsfMb7wL1RJOfA=",
+        "lastModified": 1720535198,
+        "narHash": "sha256-zwVvxrdIzralnSbcpghA92tWu2DV2lwv89xZc8MTrbg=",
         "owner": "NixOS",
         "repo": "nixpkgs",
-        "rev": "9ddcaffecdf098822d944d4147dd8da30b4e6843",
+        "rev": "205fd4226592cc83fd4c0885a3e4c9c400efabb5",
         "type": "github"
       },
       "original": {
@@ -84,11 +84,11 @@
         "treefmt-nix": "treefmt-nix"
       },
       "locked": {
-        "lastModified": 1715251496,
-        "narHash": "sha256-vRBfJCKvJtu5sYev56XStirA3lAOPv0EkoEV2Nfc+tc=",
+        "lastModified": 1724417163,
+        "narHash": "sha256-gD0N0pnKxWJcKtbetlkKOIumS0Zovgxx/nMfOIJIzoI=",
         "owner": "nix-community",
         "repo": "poetry2nix",
-        "rev": "291a863e866972f356967d0a270b259f46bf987f",
+        "rev": "7619e43c2b48c29e24b88a415256f09df96ec276",
         "type": "github"
       },
       "original": {
@@ -156,11 +156,11 @@
         ]
       },
       "locked": {
-        "lastModified": 1714058656,
-        "narHash": "sha256-Qv4RBm4LKuO4fNOfx9wl40W2rBbv5u5m+whxRYUMiaA=",
+        "lastModified": 1719749022,
+        "narHash": "sha256-ddPKHcqaKCIFSFc/cvxS14goUhCOAwsM1PbMr0ZtHMg=",
         "owner": "numtide",
         "repo": "treefmt-nix",
-        "rev": "c6aaf729f34a36c445618580a9f95a48f5e4e03f",
+        "rev": "8df5ff62195d4e67e2264df0b7f5e8c9995fd0bd",
         "type": "github"
       },
       "original": {

poetry.lock

@@ -248,22 +248,22 @@ test-randomorder = ["pytest-randomly"]
 
 [[package]]
 name = "discord-py"
-version = "2.4.0"
+version = "2.3.2"
 description = "A Python wrapper for the Discord API"
 optional = false
-python-versions = ">=3.8"
+python-versions = ">=3.8.0"
 files = [
-    {file = "discord.py-2.4.0-py3-none-any.whl", hash = "sha256:b8af6711c70f7e62160bfbecb55be699b5cb69d007426759ab8ab06b1bd77d1d"},
-    {file = "discord_py-2.4.0.tar.gz", hash = "sha256:d07cb2a223a185873a1d0ee78b9faa9597e45b3f6186df21a95cec1e9bcdc9a5"},
+    {file = "discord.py-2.3.2-py3-none-any.whl", hash = "sha256:9da4679fc3cb10c64b388284700dc998663e0e57328283bbfcfc2525ec5960a6"},
+    {file = "discord.py-2.3.2.tar.gz", hash = "sha256:4560f70f2eddba7e83370ecebd237ac09fbb4980dc66507482b0c0e5b8f76b9c"},
 ]
 
 [package.dependencies]
 aiohttp = ">=3.7.4,<4"
 
 [package.extras]
-docs = ["sphinx (==4.4.0)", "sphinx-inline-tabs (==2023.4.21)", "sphinxcontrib-applehelp (==1.0.4)", "sphinxcontrib-devhelp (==1.0.2)", "sphinxcontrib-htmlhelp (==2.0.1)", "sphinxcontrib-jsmath (==1.0.1)", "sphinxcontrib-qthelp (==1.0.3)", "sphinxcontrib-serializinghtml (==1.1.5)", "sphinxcontrib-trio (==1.1.2)", "sphinxcontrib-websupport (==1.2.4)", "typing-extensions (>=4.3,<5)"]
+docs = ["sphinx (==4.4.0)", "sphinxcontrib-trio (==1.1.2)", "sphinxcontrib-websupport", "typing-extensions (>=4.3,<5)"]
 speed = ["Brotli", "aiodns (>=1.1)", "cchardet (==2.1.7)", "orjson (>=3.5.4)"]
-test = ["coverage[toml]", "pytest", "pytest-asyncio", "pytest-cov", "pytest-mock", "typing-extensions (>=4.3,<5)", "tzdata"]
+test = ["coverage[toml]", "pytest", "pytest-asyncio", "pytest-cov", "pytest-mock", "typing-extensions (>=4.3,<5)"]
 voice = ["PyNaCl (>=1.3.0,<1.6)"]
 
 [[package]]
@@ -683,4 +683,4 @@ multidict = ">=4.0"
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.11"
-content-hash = "a69a433f44d37a63825a39563d8bf3fedea9a3c64c1a4214f54c86b1f3ddaff4"
+content-hash = "1453fcd147cbf31437ecaa4b37deb0f5d880579a2b93b5f668c28c154ed299d8"

pyproject.toml

@@ -14,7 +14,7 @@ humanize = "^4.8.0"
 parsedatetime = "^2.6"
 aiohttp = "^3.9.3"
 gidgethub = "^5.3.0"
-"discord.py" = "^2.4.0"
+"discord.py" = "^2.3.2"
 
 [tool.poetry.dev-dependencies]
 

robocop_ng/cogs/logfilereader.py

@@ -1,5 +1,6 @@
 import logging
 import re
+from typing import Optional
 
 import aiohttp
 from discord import Colour, Embed, Message, Attachment
@@ -23,6 +24,12 @@ from robocop_ng.helpers.disabled_ids import (
     remove_disabled_ro_section,
     remove_disable_id,
 )
+from robocop_ng.helpers.disabled_paths import (
+    is_path_disabled,
+    get_disabled_paths,
+    add_disabled_path,
+    remove_disabled_path,
+)
 from robocop_ng.helpers.ryujinx_log_analyser import LogAnalyser
 
 logging.basicConfig(
@@ -92,6 +99,15 @@ class LogFileReader(Cog):
                 return True
         return is_ro_section_disabled(self.bot, main_ro_section)
 
+    def contains_blocked_paths(self, log_file: str) -> Optional[str]:
+        filepaths = LogAnalyser.get_filepaths(log_file)
+        if filepaths is None:
+            return None
+        for filepath in filepaths:
+            if is_path_disabled(self.bot, filepath):
+                return filepath
+        return None
+
     async def blocked_game_action(self, message: Message) -> Embed:
         warn_command = self.bot.get_command("warn")
         if warn_command is not None:
@@ -106,7 +122,7 @@ class LogFileReader(Cog):
             )
         else:
             logging.error(
-                f"Couldn't find 'warn' command. Unable to warn {message.author}."
+                f"Couldn't find 'warn' command. Unable to warn {message.author} for uploading a log of a blocked game."
             )
 
         pirate_role = message.guild.get_role(self.bot.config.named_roles["pirate"])
@@ -116,7 +132,38 @@ class LogFileReader(Cog):
             title="⛔ Blocked game detected ⛔",
             colour=Colour(0xFF0000),
             description="This log contains a blocked game and has been removed.\n"
-            "The user has been warned and the pirate role was applied.",
+            "The user has been warned and the pirate role has been applied.",
+        )
+        embed.set_footer(text=f"Log uploaded by @{message.author.name}")
+        await message.delete()
+        return embed
+
+    async def blocked_path_action(self, message: Message, blocked_path: str) -> Embed:
+        warn_command = self.bot.get_command("warn")
+        if warn_command is not None:
+            warn_message = await message.reply(
+                ".warn This log contains blocked content in paths."
+            )
+            warn_context = await self.bot.get_context(warn_message)
+            await warn_context.invoke(
+                warn_command,
+                target=None,
+                reason=f"This log contains blocked content in paths: '{blocked_path}'",
+            )
+        else:
+            logging.error(
+                f"Couldn't find 'warn' command. Unable to warn {message.author} for uploading a log "
+                f"containing a blocked content in paths."
+            )
+
+        pirate_role = message.guild.get_role(self.bot.config.named_roles["pirate"])
+        await message.author.add_roles(pirate_role)
+
+        embed = Embed(
+            title="⛔ Blocked content in path detected ⛔",
+            colour=Colour(0xFF0000),
+            description="This log contains paths containing blocked content and has been removed.\n"
+            "The user has been warned and the pirate role has been applied.",
         )
         embed.set_footer(text=f"Log uploaded by @{message.author.name}")
         await message.delete()
@@ -243,6 +290,9 @@ class LogFileReader(Cog):
 
         if self.is_game_blocked(log_file):
             return await self.blocked_game_action(message)
+        blocked_path = self.contains_blocked_paths(log_file)
+        if blocked_path:
+            return await self.blocked_path_action(message, blocked_path)
 
         for role in message.author.roles:
             if role.id in self.disallowed_roles:
@@ -449,6 +499,61 @@ class LogFileReader(Cog):
         else:
             return await ctx.send(f"No read-only section blocked for '{disable_id}'.")
 
+    @commands.check(check_if_staff)
+    @commands.command(
+        aliases=["disallow_path", "forbid_path", "block_path", "blockpath"]
+    )
+    async def disable_path(self, ctx: Context, block_path: str):
+        if add_disabled_path(self.bot, block_path):
+            return await ctx.send(f"Path content `{block_path}` is now blocked!")
+        else:
+            return await ctx.send(f"Path content `{block_path}` is already blocked.")
+
+    @commands.check(check_if_staff)
+    @commands.command(
+        aliases=[
+            "allow_path",
+            "unblock_path",
+            "unblockpath",
+        ]
+    )
+    async def enable_path(self, ctx: Context, block_path: str):
+        if remove_disabled_path(self.bot, block_path):
+            return await ctx.send(f"Path content `{block_path}` is now unblocked!")
+        else:
+            return await ctx.send(f"No blocked path content '{block_path}' found.")
+
+    @commands.check(check_if_staff)
+    @commands.command(
+        aliases=[
+            "disabled_paths",
+            "blocked_paths",
+            "listdisabledpaths",
+            "listblockedpaths",
+            "list_blocked_paths",
+        ]
+    )
+    async def list_disabled_paths(self, ctx: Context):
+        messages = []
+        disabled_paths = get_disabled_paths(self.bot)
+
+        message = (
+            "**Blocking analysis of logs containing the following content in paths:**\n"
+        )
+        for entry in disabled_paths:
+            if len(message) >= 1500:
+                messages.append(message)
+                message = f"- `{entry}`\n"
+            else:
+                message += f"- `{entry}`\n"
+
+        if message not in messages:
+            # Add the last message as well
+            messages.append(message)
+
+        for msg in messages:
+            await ctx.send(msg)
+
     async def analyse_log_message(self, message: Message, attachment_index=0):
         author_id = message.author.id
         author_mention = message.author.mention
@@ -563,6 +668,12 @@ class LogFileReader(Cog):
                         return await message.channel.send(
                             content=None, embed=await self.blocked_game_action(message)
                         )
+                    blocked_path = self.contains_blocked_paths(log_file)
+                    if blocked_path:
+                        return await message.channel.send(
+                            content=None,
+                            embed=await self.blocked_path_action(message, blocked_path),
+                        )
             elif (
                 is_log_file
                 and is_ryujinx_log_file

robocop_ng/helpers/disabled_paths.py

@@ -0,0 +1,47 @@
+import json
+import os
+
+from robocop_ng.helpers.data_loader import read_json
+
+
+def get_disabled_paths_path(bot) -> str:
+    return os.path.join(bot.state_dir, "data/disabled_paths.json")
+
+
+def get_disabled_paths(bot) -> list[str]:
+    disabled_paths = read_json(bot, get_disabled_paths_path(bot))
+    if "paths" not in disabled_paths.keys():
+        return []
+    return disabled_paths["paths"]
+
+
+def set_disabled_paths(bot, contents: list[str]):
+    with open(get_disabled_paths_path(bot), "w") as f:
+        json.dump({"paths": contents}, f)
+
+
+def is_path_disabled(bot, path: str) -> bool:
+    for disabled_path in get_disabled_paths(bot):
+        if disabled_path in path.strip().lower():
+            return True
+    return False
+
+
+def add_disabled_path(bot, disabled_path: str) -> bool:
+    disabled_path = disabled_path.strip().lower()
+    disabled_paths = get_disabled_paths(bot)
+    if disabled_path not in disabled_paths:
+        disabled_paths.append(disabled_path)
+        set_disabled_paths(bot, disabled_paths)
+        return True
+    return False
+
+
+def remove_disabled_path(bot, disabled_path: str) -> bool:
+    disabled_path = disabled_path.strip().lower()
+    disabled_paths = get_disabled_paths(bot)
+    if disabled_path in disabled_paths:
+        disabled_paths.remove(disabled_path)
+        set_disabled_paths(bot, disabled_paths)
+        return True
+    return False

robocop_ng/helpers/ryujinx_log_analyser.py

@@ -41,6 +41,13 @@ class LogAnalyser:
             re.search("Load.*Application: Loading as [Hh]omebrew", log_file) is not None
         )
 
+    @staticmethod
+    def get_filepaths(log_file: str) -> set[str]:
+        return set(
+            x.rstrip("\u0000")
+            for x in re.findall(r"(?:[A-Za-z]:)?(?:[\\/]+[^\\/:\"\r\n]+)+", log_file)
+        )
+
     @staticmethod
     def get_main_ro_section(log_file: str) -> Optional[dict[str, str]]:
         ro_section_matches = re.findall(
@@ -81,7 +88,7 @@ class LogAnalyser:
             else:
                 app_id = ""
             bids_match_all = re.findall(
-                r"Build ids found for title ([a-zA-Z0-9]*):[\n\r]*((?:\s+.*[\n\r]+)+)",
+                r"Build ids found for (?:title|application) ([a-zA-Z0-9]*):[\n\r]*((?:\s+.*[\n\r]+)+)",
                 log_file,
             )
             if bids_match_all and len(bids_match_all) > 0:
@@ -443,6 +450,10 @@ class LogAnalyser:
                     "**⚠️ AMD GPU users should consider using Vulkan graphics backend**"
                 )
 
+    def __get_cpu_notes(self):
+        if "VirtualApple" in self._hardware_info["cpu"]:
+            self._notes.add("🔴 **Rosetta should be disabled**")
+
     def __get_log_notes(self):
         default_logs = ["Info", "Warning", "Error", "Guest"]
         user_logs = []
@@ -572,6 +583,7 @@ class LogAnalyser:
 
         self.__get_controller_notes()
         self.__get_os_notes()
+        self.__get_cpu_notes()
 
         if (
             self._emu_info["ryu_firmware"] == "Unknown"
@@ -706,6 +718,7 @@ class LogAnalyser:
             "errors": self._log_errors,
             "settings": self._settings,
             "app_info": self.get_app_info(self._log_text),
+            "paths": list(self.get_filepaths(self._log_text)),
         }