rename vendor dir to pkgs

archivebox/pkgs/abx-plugin-ripgrep-search/abx_plugin_ripgrep_search/__init__.py

@@ -0,0 +1,31 @@
+__package__ = 'abx_plugin_ripgrep_search'
+__label__ = 'Ripgrep Search'
+__homepage__ = 'https://github.com/BurntSushi/ripgrep'
+
+import abx
+
+@abx.hookimpl
+def get_CONFIG():
+    from .config import RIPGREP_CONFIG
+    
+    return {
+        'RIPGREP_CONFIG': RIPGREP_CONFIG
+    }
+
+
+@abx.hookimpl
+def get_BINARIES():
+    from .binaries import RIPGREP_BINARY
+    
+    return {
+        'ripgrep': RIPGREP_BINARY
+    }
+
+
+@abx.hookimpl
+def get_SEARCHBACKENDS():
+    from .searchbackend import RIPGREP_SEARCH_BACKEND
+    
+    return {
+        'ripgrep': RIPGREP_SEARCH_BACKEND,
+    }

archivebox/pkgs/abx-plugin-ripgrep-search/abx_plugin_ripgrep_search/binaries.py

@@ -0,0 +1,23 @@
+__package__ = 'abx_plugin_ripgrep_search'
+
+from typing import List
+
+from pydantic import InstanceOf
+from pydantic_pkgr import BinProvider, BinaryOverrides, BinName, Binary
+
+from abx_plugin_default_binproviders import apt, brew, env
+
+
+from .config import RIPGREP_CONFIG
+
+
+class RipgrepBinary(Binary):
+    name: BinName = RIPGREP_CONFIG.RIPGREP_BINARY
+    binproviders_supported: List[InstanceOf[BinProvider]] = [apt, brew, env]
+
+    overrides: BinaryOverrides = {
+        apt.name: {'packages': ['ripgrep']},
+        brew.name: {'packages': ['ripgrep']},
+    }
+
+RIPGREP_BINARY = RipgrepBinary()

archivebox/pkgs/abx-plugin-ripgrep-search/abx_plugin_ripgrep_search/config.py

@@ -0,0 +1,29 @@
+__package__ = 'abx_plugin_ripgrep_search'
+
+from pathlib import Path
+from typing import List
+
+from pydantic import Field
+
+from abx_spec_config.base_configset import BaseConfigSet
+
+from archivebox.config import CONSTANTS
+from archivebox.config.common import SEARCH_BACKEND_CONFIG
+
+
+class RipgrepConfig(BaseConfigSet):
+    RIPGREP_BINARY: str = Field(default='rg')
+    
+    RIPGREP_IGNORE_EXTENSIONS: str = Field(default='css,js,orig,svg')
+    RIPGREP_ARGS_DEFAULT: List[str] = Field(default=lambda c: [
+        # https://github.com/BurntSushi/ripgrep/blob/master/GUIDE.md
+        f'--type-add=ignore:*.{{{c.RIPGREP_IGNORE_EXTENSIONS}}}',
+        '--type-not=ignore',
+        '--ignore-case',
+        '--files-with-matches',
+        '--regexp',
+    ])
+    RIPGREP_SEARCH_DIR: Path = CONSTANTS.ARCHIVE_DIR
+    RIPGREP_TIMEOUT: int = Field(default=lambda: SEARCH_BACKEND_CONFIG.SEARCH_BACKEND_TIMEOUT)
+
+RIPGREP_CONFIG = RipgrepConfig()

archivebox/pkgs/abx-plugin-ripgrep-search/abx_plugin_ripgrep_search/searchbackend.py

@@ -0,0 +1,55 @@
+__package__ = 'abx_plugin_ripgrep_search'
+
+import re
+import subprocess
+
+from typing import List, Iterable
+
+from abx_spec_searchbackend import BaseSearchBackend
+
+from .binaries import RIPGREP_BINARY
+from .config import RIPGREP_CONFIG
+
+
+
+# regex to match archive/<ts>/... snapshot dir names
+TIMESTAMP_REGEX =  re.compile(r'\/([\d]+\.[\d]+)\/')
+
+class RipgrepSearchBackend(BaseSearchBackend):
+    name: str = 'ripgrep'
+    docs_url: str = 'https://github.com/BurntSushi/ripgrep'
+    
+    @staticmethod
+    def index(snapshot_id: str, texts: List[str]):
+        return
+
+    @staticmethod
+    def flush(snapshot_ids: Iterable[str]):
+        return
+
+    @staticmethod
+    def search(text: str) -> List[str]:
+        from core.models import Snapshot
+        
+        ripgrep_binary = RIPGREP_BINARY.load()
+        if not ripgrep_binary.version:
+            raise Exception("ripgrep binary not found, install ripgrep to use this search backend")
+    
+        cmd = [
+            ripgrep_binary.abspath, 
+            *RIPGREP_CONFIG.RIPGREP_ARGS_DEFAULT,
+            text,
+            str(RIPGREP_CONFIG.RIPGREP_SEARCH_DIR),
+        ]
+        proc = subprocess.run(cmd, timeout=RIPGREP_CONFIG.RIPGREP_TIMEOUT, capture_output=True, text=True)
+        timestamps = set()
+        for path in proc.stdout.splitlines():
+            ts = TIMESTAMP_REGEX.findall(path)
+            if ts:
+                timestamps.add(ts[0])
+        
+        snap_ids = [str(id) for id in Snapshot.objects.filter(timestamp__in=timestamps).values_list('pk', flat=True)]
+    
+        return snap_ids
+
+RIPGREP_SEARCH_BACKEND = RipgrepSearchBackend()

archivebox/pkgs/abx-plugin-ripgrep-search/pyproject.toml

@@ -0,0 +1,18 @@
+[project]
+name = "abx-plugin-ripgrep-search"
+version = "2024.10.28"
+description = "Add your description here"
+readme = "README.md"
+requires-python = ">=3.10"
+dependencies = [
+    "abx>=0.1.0",
+    "abx-spec-config>=0.1.0",
+    "abx-spec-searchbackend>=0.1.0",
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project.entry-points.abx]
+abx_plugin_ripgrep_search = "abx_plugin_ripgrep_search"