diff --git a/pygitweb/README.md b/pygitweb/README.md index 8be1ee1..8e3dcf1 100644 --- a/pygitweb/README.md +++ b/pygitweb/README.md @@ -78,7 +78,8 @@ Load the `/docs` page for a detailed view of routes (below the readme) if in deb | blame_data | — | | TODO | | rss | — | | TODO | | atom | — | | TODO | -| search | — | | TODO | +| search | `patterns`, `paths`, `globs`, `heading`, `sort`, `max_count`, `multiline` | `GET /project/{project}?a=search&patterns=...` | Ripgrep search (via `python-ripgrep`); returns JSON. `paths` are relative to the project tree and cannot escape it. | +| search (page) | — | `GET /project/{project}/search` | Per-project search UI page. | | search_help | — | | TODO | ## Query parameter short names (CGI mapping) diff --git a/pygitweb/actions.py b/pygitweb/actions.py index 4f0dfb2..964e830 100644 --- a/pygitweb/actions.py +++ b/pygitweb/actions.py @@ -9,12 +9,15 @@ import base64 import mimetypes import os from datetime import UTC, datetime, timedelta, timezone +from pathlib import Path from typing import Any, TypedDict from urllib.parse import quote, urlencode import pygit2 from fastapi import HTTPException, Request -from fastapi.responses import HTMLResponse, PlainTextResponse, Response +from fastapi.responses import HTMLResponse, JSONResponse, PlainTextResponse, Response +from python_ripgrep import PySortMode, PySortModeKind +from python_ripgrep import search as rg_search from pygitweb.config import BLOB_LANG, settings from pygitweb.formatting import age_string, sanitize, to_utf8 @@ -41,6 +44,77 @@ from pygitweb.templates_env import POSTAMBLE, PREAMBLE, env, jinja_escape from pygitweb.validation import is_valid_pathname, is_valid_ref_format +def _split_query_list(values: list[str]) -> list[str]: + parts: list[str] = [] + for v in values: + for seg in (v or "").split(","): + s = seg.strip() + if s: + parts.append(s) + return parts + + +def _parse_query_bool(value: str | None) -> bool | None: + if value is None: + return None + v = value.strip().lower() + if v in ("1", "true", "yes", "on"): + return True + if v in ("0", "false", "no", "off"): + return False + raise HTTPException(status_code=400, detail=f"Invalid boolean: {value!r}") + + +def _project_worktree_root(project: str) -> Path: + root = (Path(settings.PROJECTROOT) / project).resolve() + if not root.exists() or not root.is_dir(): + raise HTTPException(status_code=404, detail="Project directory not found") + return root + + +def _validated_search_paths(project_root: Path, raw_paths: list[str]) -> list[str]: + root = project_root.resolve() + if not raw_paths: + return [str(root)] + + paths: list[str] = [] + for raw in raw_paths: + p = (raw or "").strip() + if not p: + continue + candidate = Path(p) + if candidate.is_absolute() or ":" in p: + raise HTTPException(status_code=400, detail="path not relative to the project root") + abs_path = (root / p).resolve() + if not abs_path.is_relative_to(root): + raise HTTPException(status_code=400, detail="path not within the project root") + if not abs_path.exists(): + raise HTTPException(status_code=400, detail=f"path entry does not exist: {p}") + paths.append(str(abs_path)) + + if not paths: + return [str(root)] + return paths + + +_SORT_KINDS: tuple[str, ...] = ("Path", "LastModified", "LastAccessed", "Created") + + +def _parse_search_sort(value: str | None) -> PySortMode | None: + if value is None or not value.strip(): + return None + raw = value.strip() + reverse = False + if raw.startswith("-"): + reverse = True + raw = raw[1:].strip() + if raw not in _SORT_KINDS: + allowed = ", ".join(_SORT_KINDS) + raise HTTPException(status_code=400, detail=f"Invalid sort kind; allowed: {allowed}") + kind = getattr(PySortModeKind, raw) + return PySortMode(kind=kind, reverse=reverse) + + def _pagination_url(request: Request, page: int, pagecount: int) -> str: """Build URL for a pagination page, preserving path and other query params.""" params = dict(request.query_params) @@ -336,6 +410,80 @@ def git_object(project: str, h: str | None) -> Response: raise HTTPException(status_code=404, detail="Unknown object type") +def git_search(project: str, request: Request) -> JSONResponse: + qp = request.query_params + patterns = _split_query_list(qp.getlist("patterns")) + if not patterns: + raise HTTPException(status_code=400, detail="patterns is required (repeat ?patterns= or use comma-separated)") + + project_root = _project_worktree_root(project) + paths = _validated_search_paths(project_root, _split_query_list(qp.getlist("paths"))) + + globs = _split_query_list(qp.getlist("globs")) + heading = _parse_query_bool(qp.get("heading")) + multiline = _parse_query_bool(qp.get("multiline")) + sort = _parse_search_sort(qp.get("sort")) + + max_count_raw = qp.get("max_count") + max_count: int = 25 + if max_count_raw is not None and max_count_raw.strip() != "": + try: + max_count = int(max_count_raw) + except ValueError as e: + raise HTTPException(status_code=400, detail="max_count must be an integer") from e + if max_count < 1: + raise HTTPException(status_code=400, detail="max_count must be at least 1") + + try: + results = rg_search( + patterns=patterns, + paths=paths, + globs=(globs or None), + heading=heading, + sort=sort, + max_count=max_count, + multiline=multiline, + ) + except Exception as e: + raise HTTPException(status_code=500, detail=f"Search failed: {e}") from e + + root_str = str(project_root.resolve()) + rel_paths: list[str] = [] + for p in paths: + abs_p = str(Path(p).resolve()) + rel = os.path.relpath(abs_p, root_str) + rel_paths.append("" if rel == "." else rel.replace(os.sep, "/")) + + return JSONResponse({ + "project": project, + "project_root": root_str, + "patterns": patterns, + "paths": rel_paths, + "globs": globs, + "heading": heading, + "sort": (qp.get("sort") or None), + "max_count": max_count, + "multiline": multiline, + "results": results, + }) + + +def git_search_page(project: str) -> HTMLResponse: + """Per-project search UI page. Calls /project/{project}?a=search via fetch.""" + project_enc = quote(project, safe="/") + pre = PREAMBLE.render( + title=f"Search - {project}", + site_name=settings.SITE_NAME, + ) + body = env.get_template("search.html").render( + project=project, + project_enc=project_enc, + project_url=f"/project/{project_enc}", + search_api_url=f"/project/{project_enc}", + ) + return HTMLResponse(f"{pre}{body}{POSTAMBLE}") + + def git_summary(project: str, extra_rows: list[list[str]] | None = None) -> HTMLResponse: """Project summary page. Port of git_summary.""" descr = git_get_project_description(project) or "" @@ -376,6 +524,7 @@ def git_summary(project: str, extra_rows: list[list[str]] | None = None) -> HTML ["Tags", f"view tags"], ["Remotes", f"view remotes"], ["Board", board_value], + ["Search", f"search files"], ] if extra_rows: rows.extend(extra_rows) diff --git a/pygitweb/main.py b/pygitweb/main.py index a90a9ff..107c750 100644 --- a/pygitweb/main.py +++ b/pygitweb/main.py @@ -31,6 +31,8 @@ from pygitweb.actions import ( git_patch, git_patches, git_remotes, + git_search, + git_search_page, git_shortlog, git_summary, git_tag, @@ -513,6 +515,13 @@ def project_summary_ref_state( return summary_ref_state(project, ref) +@app.get("/project/{project:path}/search", response_class=HTMLResponse) +def project_search_page(project: str) -> HTMLResponse: + """Per-project search page (UI only; results fetched from /project/{project}?a=search).""" + _validate_project(project) + return git_search_page(project) + + @app.get("/project/{project:path}/subpage/{subpage_name}", response_class=HTMLResponse) def project_subpage( project: str, @@ -635,6 +644,8 @@ def dispatch( return git_remotes(project) if action == "object": return git_object(project, hash_param) + if action == "search": + return git_search(project, request) # Stub others with minimal response pre = PREAMBLE.render(title=f"{action} - {project}", site_name=settings.SITE_NAME) return HTMLResponse(f"{pre}

Action: {jinja_escape(action)}

Project: {jinja_escape(project)}

{POSTAMBLE}") diff --git a/pygitweb/pyproject.toml b/pygitweb/pyproject.toml index cf30b5c..7b6ec90 100644 --- a/pygitweb/pyproject.toml +++ b/pygitweb/pyproject.toml @@ -17,6 +17,7 @@ dependencies = [ "orjson>=0.19.0", "pygit2>=1.12.0", "pydantic-settings>=2.13.0", + "python-ripgrep==0.0.9", ] [tool.setuptools] diff --git a/pygitweb/static/main.css b/pygitweb/static/main.css index 5faf3f8..4725a0f 100644 --- a/pygitweb/static/main.css +++ b/pygitweb/static/main.css @@ -840,6 +840,42 @@ tbody tr { } } +/* Search page: input, status, and result lines */ +.search-status { + font-style: italic; +} + +.search-results { + display: flex; + flex-direction: column; + gap: 0.25rem; + + &.is-loading { + opacity: 0.6; + } +} + +.search-result { + background-color: var(--pgw-bg-raised); + border: 1px solid var(--pgw-border); + border-radius: 0.25rem; + padding: 0.25rem 0.5rem; + + &:hover { + background-color: var(--pgw-table-hover); + } +} + +.search-result-line { + margin: 0; + font-family: ui-monospace, "Fira Code", "Fira Mono", monospace; + font-size: 0.85rem; + line-height: 1.4; + color: var(--pgw-text); + white-space: pre-wrap; + word-break: break-word; +} + .subpage-frame { display: block; width: 100%; diff --git a/pygitweb/static/search-page.js b/pygitweb/static/search-page.js new file mode 100644 index 0000000..ec952c5 --- /dev/null +++ b/pygitweb/static/search-page.js @@ -0,0 +1,139 @@ +(function () { + "use strict"; + + const apiUrl = window.SEARCH_API_URL; + const elPattern = document.getElementById("search-patterns"); + const elPaths = document.getElementById("search-paths"); + const elGlobs = document.getElementById("search-globs"); + const elSort = document.getElementById("search-sort"); + const elMax = document.getElementById("search-max-count"); + const elHead = document.getElementById("search-heading"); + const elMulti = document.getElementById("search-multiline"); + const elStatus = document.getElementById("search-status"); + const elResults = document.getElementById("search-results"); + const elAdvanced = document.getElementById("search-advanced"); + + function escapeHtml(s) { + return String(s).replace(/[&<>"']/g, function (c) { + return ({ "&": "&", "<": "<", ">": ">", '"': """, "'": "'" })[c]; + }); + } + + function readUrl() { + const p = new URLSearchParams(window.location.search); + if (p.has("patterns")) elPattern.value = p.get("patterns"); + if (p.has("paths")) elPaths.value = p.get("paths"); + if (p.has("globs")) elGlobs.value = p.get("globs"); + if (p.has("sort")) elSort.value = p.get("sort"); + elMax.value = p.has("max_count") ? p.get("max_count") : "25"; + if (p.has("heading")) elHead.checked = p.get("heading") === "true"; + if (p.has("multiline")) elMulti.checked = p.get("multiline") === "true"; + const anyAdvanced = ["paths", "globs", "sort", "multiline"].some((k) => p.has(k)); + if (anyAdvanced) elAdvanced.open = true; + } + + function buildParams() { + const qs = new URLSearchParams(); + qs.set("a", "search"); + const pat = elPattern.value.trim(); + if (pat) qs.set("patterns", pat); + const paths = elPaths.value.trim(); + if (paths) qs.set("paths", paths); + const globs = elGlobs.value.trim(); + if (globs) qs.set("globs", globs); + if (elSort.value) qs.set("sort", elSort.value); + const max = elMax.value.trim(); + if (max) qs.set("max_count", max); + qs.set("heading", elHead.checked ? "true" : "false"); + qs.set("multiline", elMulti.checked ? "true" : "false"); + return qs; + } + + function persistUrl(qs) { + const url = new URL(window.location.href); + const keep = new Set(Array.from(qs.keys()).filter((k) => k !== "a")); + Array.from(url.searchParams.keys()).forEach((k) => { + if (!keep.has(k)) url.searchParams.delete(k); + }); + for (const [k, v] of qs.entries()) { + if (k === "a") continue; + url.searchParams.set(k, v); + } + window.history.replaceState(null, "", url); + } + + function renderResults(data) { + const lines = Array.isArray(data.results) ? data.results : []; + if (lines.length === 0) { + elResults.innerHTML = ""; + return; + } + const items = lines.map((line) => { + return '
' + escapeHtml(line) + "
"; + }); + elResults.innerHTML = items.join(""); + } + + let pendingCtrl = null; + + async function runSearch() { + const pat = elPattern.value.trim(); + if (!pat) { + elStatus.textContent = ""; + elResults.innerHTML = ""; + persistUrl(new URLSearchParams()); + return; + } + const qs = buildParams(); + persistUrl(qs); + if (pendingCtrl) pendingCtrl.abort(); + pendingCtrl = new AbortController(); + elStatus.textContent = "Searching\u2026"; + elResults.classList.add("is-loading"); + try { + const resp = await fetch(apiUrl + "?" + qs.toString(), { + signal: pendingCtrl.signal, + headers: { Accept: "application/json" }, + }); + if (!resp.ok) { + let detail = resp.statusText; + try { + const j = await resp.json(); + if (j && j.detail) detail = j.detail; + } catch (_) { /* ignore */ } + elStatus.textContent = "Error " + resp.status + ": " + detail; + elResults.innerHTML = ""; + return; + } + const data = await resp.json(); + const n = (data.results || []).length; + elStatus.textContent = n + " match" + (n === 1 ? "" : "es"); + renderResults(data); + } catch (err) { + if (err && err.name === "AbortError") return; + elStatus.textContent = "Error: " + (err && err.message ? err.message : String(err)); + } finally { + elResults.classList.remove("is-loading"); + } + } + + let debounceTimer = null; + function debounced() { + clearTimeout(debounceTimer); + debounceTimer = setTimeout(runSearch, 500); + } + + for (const el of [elPattern, elPaths, elGlobs, elMax]) el.addEventListener("input", debounced); + for (const el of [elSort, elHead, elMulti]) el.addEventListener("change", debounced); + + elPattern.addEventListener("keydown", function (event) { + if (event.key === "Enter") { + event.preventDefault(); + clearTimeout(debounceTimer); + runSearch(); + } + }); + + readUrl(); + runSearch(); +})(); diff --git a/pygitweb/templates/search.html b/pygitweb/templates/search.html new file mode 100644 index 0000000..687d18a --- /dev/null +++ b/pygitweb/templates/search.html @@ -0,0 +1,77 @@ +{# Search UI page. Expects: project, project_enc, project_url, search_api_url. #} + +

+ {{ project }} + / + search +

+ +
+
+
+ + +
+ Advanced options +
+
+ + +
Restricted to the project tree.
+
+
+ + +
+
+ + +
+
+ + +
+
+ +
+
+ +
+
+
+
+
+
+ +
+
+ + + diff --git a/uv.lock b/uv.lock index a654d5e..db25484 100644 --- a/uv.lock +++ b/uv.lock @@ -957,6 +957,7 @@ dependencies = [ { name = "pygit2" }, { name = "pygittools" }, { name = "python-multipart" }, + { name = "python-ripgrep" }, { name = "uvicorn", extra = ["standard"] }, ] @@ -969,6 +970,7 @@ requires-dist = [ { name = "pygit2", specifier = ">=1.12.0" }, { name = "pygittools", editable = "pygittools" }, { name = "python-multipart", specifier = ">=0.0.6" }, + { name = "python-ripgrep", specifier = "==0.0.9" }, { name = "uvicorn", extras = ["standard"], specifier = ">=0.24.0" }, ] @@ -1074,6 +1076,28 @@ wheels = [ ] [[package]] +name = "python-ripgrep" +version = "0.0.9" +source = { registry = "https://pypi.org/simple" } +sdist = { url = "https://files.pythonhosted.org/packages/e1/8f/98b524a7186e722dd5c33e033db28e30a0000438c5b2493b842f2766bcd0/python_ripgrep-0.0.9.tar.gz", hash = "sha256:5a65b8252c76df9c523b837c2a4f86478cc2af996bac4d10f9b321c3768e259b", size = 45207, upload-time = "2025-07-30T19:03:29.361Z" } +wheels = [ + { url = "https://files.pythonhosted.org/packages/75/38/ef6ae18f59ee03e4e6666f86ccd1b7a31ee503768e6497eb71dfac1acb3d/python_ripgrep-0.0.9-cp38-abi3-macosx_10_12_x86_64.whl", hash = "sha256:a1a38b5a39beeb1b07a7e0d27942442fd61e61065660aa986051c90bcdaa323f", size = 1607445, upload-time = "2025-07-30T19:56:25.056Z" }, + { url = "https://files.pythonhosted.org/packages/eb/1a/3d20679e64dacc467d95bd6630d76e1b57125efa09cb6a5f9e158cdb1f6e/python_ripgrep-0.0.9-cp38-abi3-macosx_11_0_arm64.whl", hash = "sha256:83d9b7908ce61a2fc3bb4e2318705c87099cc43a4453e6768e860a329d0a1a7a", size = 1543925, upload-time = "2025-07-30T19:03:28.187Z" }, + { url = "https://files.pythonhosted.org/packages/f6/50/77f6eb5dfe56d55fd42da383017a3c2a9d42555bea232fc7ebb8d3291271/python_ripgrep-0.0.9-cp38-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2e11004a12a31247c1d5a77bfb8a71793fa552a355c96498f68937348175ba2e", size = 1719093, upload-time = "2025-07-30T19:56:16.426Z" }, + { url = "https://files.pythonhosted.org/packages/9e/0f/4f2f624bc30783018649779c8b669e37d866c5b5caeb7e3a8e53bcc6256b/python_ripgrep-0.0.9-cp38-abi3-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:850ac515eb2ff370ccb0ddd934ca32f06f75e5fa5a3099111dad140079a26d99", size = 1669712, upload-time = "2025-07-30T19:56:18.154Z" }, + { url = "https://files.pythonhosted.org/packages/25/b5/5fa748faeaee82e65983e450b5b52e626f0bdcb43420e113c3d0831c9e49/python_ripgrep-0.0.9-cp38-abi3-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:193db1978e52ec2959913bee0189a1f5366178bcf19473de702ac9235f122740", size = 1816691, upload-time = "2025-07-30T19:56:22.497Z" }, + { url = "https://files.pythonhosted.org/packages/97/05/0c2ab823a9fcbbbf3f826c30bac762fc0a520f6ee111c9a83222cf810e27/python_ripgrep-0.0.9-cp38-abi3-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:ec6a25229cc5998860ed4c4a14e1b29816ed98f8d17d4847d6fbe625e0bfdfda", size = 2032756, upload-time = "2025-07-30T19:56:19.87Z" }, + { url = "https://files.pythonhosted.org/packages/ae/3f/19784133a96609e7682cf8020068d0f4810509d27eee6916b51ebbdc441b/python_ripgrep-0.0.9-cp38-abi3-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:8f9ff586d1f6667f60e29d5eadebcf28c01f9d68851662ad84665d83d940ed2d", size = 1878113, upload-time = "2025-07-30T19:56:21.315Z" }, + { url = "https://files.pythonhosted.org/packages/74/b4/ac6b588d8437cdf22d9450cd87a2939ca91bd304ce7fa400484978ce6394/python_ripgrep-0.0.9-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9acae870b9dba3d5a70aeeab687809dfad7c2afe3c422fedb9fca3d578b5d422", size = 1780017, upload-time = "2025-07-30T19:56:23.734Z" }, + { url = "https://files.pythonhosted.org/packages/85/a5/baab3499501710421ba9ea0199f3ba2450bdf8eedaf9de6b5a043513d925/python_ripgrep-0.0.9-cp38-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:2be02995828e36c8b37cdb6af50178021bc580f0ca6766f4d14806aa904641ae", size = 1893412, upload-time = "2025-07-30T19:56:26.123Z" }, + { url = "https://files.pythonhosted.org/packages/73/88/d6d06fb08401a00ef58e6a3f6f486246e4cfd3b82cc7fed8c187fd273547/python_ripgrep-0.0.9-cp38-abi3-musllinux_1_2_armv7l.whl", hash = "sha256:fcf0007fd3e39b5fd2cc941af5eacb24de2d5268ad41ae6edaf7abbc5b3965d4", size = 1927522, upload-time = "2025-07-30T19:56:27.458Z" }, + { url = "https://files.pythonhosted.org/packages/7f/41/e8cbd79664bb091bfd3310758809c73ccf05d92a94b9768d913f667df03c/python_ripgrep-0.0.9-cp38-abi3-musllinux_1_2_i686.whl", hash = "sha256:004b1223684f457ab96d180edb2ba64bc1c0c0b26053254bc345443e6a3d83c8", size = 1915262, upload-time = "2025-07-30T19:56:28.676Z" }, + { url = "https://files.pythonhosted.org/packages/dd/98/4830184a0fba24958f68f4c0e6749dd7ace120e08cbe1a70707fc05a98c1/python_ripgrep-0.0.9-cp38-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:80cb59c90e98cf83325b4dbad36c1569fc9c658aa2d1244943bf0d98352227d4", size = 1951646, upload-time = "2025-07-30T19:56:29.741Z" }, + { url = "https://files.pythonhosted.org/packages/30/2f/de02bc9476f56ae4d0a7838319d8b02b5f3ac45a9eee710f17c49738797e/python_ripgrep-0.0.9-cp38-abi3-win32.whl", hash = "sha256:69b4039f1c86313a4f39350db19a720c41a1b5396b61abb5c9b9077794b740ad", size = 1308752, upload-time = "2025-07-30T19:56:31.999Z" }, + { url = "https://files.pythonhosted.org/packages/53/0f/bcbec2b9fb69be66600c63b97b18b50fa24decc9dd0cc826acc9d8c970dc/python_ripgrep-0.0.9-cp38-abi3-win_amd64.whl", hash = "sha256:fb92b64c9e854ab74e15ab4348f93fbc2c2a7384ab946be5e1fa59cc98cb516f", size = 1410074, upload-time = "2025-07-30T19:56:30.893Z" }, +] + +[[package]] name = "pyyaml" version = "6.0.3" source = { registry = "https://pypi.org/simple" }