From 097be623f497712d1e0583080f2a74c931c3855e Mon Sep 17 00:00:00 2001
From: Dariusz L <dl122@wp.pl>
Date: Mon, 20 Jul 2026 16:16:09 +0200
Subject: [PATCH] feat: preserve CivArchive page URLs separate from HuggingFace
 mirrors, fix HTML array escaping, and accept provider array payloads

Backend:
- Keep CivArchive source page URL distinct from HuggingFace download mirrors in download metadata
- Handle CivitAI hash lookup HTTP errors and unexpected JSON shapes without caching failures
- Allow network JSON requests to return arrays as well as objects

Frontend:
- Escape each item individually when interpolating arrays into HTML templates
- Prefer the selected provider's page URL over cross-provider mirrors in context menus and download targets
- Reuse source link resolution for model info display

Tests:
- Add coverage for CivArchive metadata URL isolation, HTML array escaping, provider array responses, and CivitAI hash lookup caching behavior
---
 core/downloader.py                            | 23 +++++
 core/network_utils.py                         | 12 ++-
 core/sources/civitai.py                       | 18 +++-
 tests/test_downloader_metadata.py             | 33 +++++++
 tests/test_downloads_tab_workflow_route.mjs   | 95 ++++++++++++++++++-
 tests/test_source_internals.py                | 49 ++++++++++
 tests/test_unified_helpers.py                 | 26 +++++
 .../search/download_target_methods.js         | 40 +++++++-
 web/resolver/utils/html_utils.js              |  4 +-
 web/resolver/views/model_info_methods.js      | 30 ++++--
 10 files changed, 311 insertions(+), 19 deletions(-)

diff --git a/core/downloader.py b/core/downloader.py
index 6c61f1a..016f670 100644
--- a/core/downloader.py
+++ b/core/downloader.py
@@ -550,6 +550,29 @@ def build_lora_manager_metadata(
         source.get("platform_url"),
         source_url,
     )
+    if source_name == "civarchive":
+        civarchive_page_url = next(
+            (
+                str(value).strip()
+                for value in (
+                    source.get("version_url"),
+                    source.get("model_url"),
+                    source.get("page_url"),
+                    source.get("source_url"),
+                    source.get("url"),
+                )
+                if value
+                and host_matches_domain(urlparse(str(value)).hostname, "civarchive.com")
+            ),
+            "",
+        )
+        if not civarchive_page_url and model_id:
+            civarchive_page_url = f"https://civarchive.com/models/{model_id}"
+            if version_id:
+                civarchive_page_url += f"?modelVersionId={version_id}"
+        if not civarchive_page_url and sha256:
+            civarchive_page_url = f"https://civarchive.com/sha256/{sha256}"
+        source_page_url = civarchive_page_url
     platform_url = _first_present(source.get("platform_url"), source.get("platformUrl"))
     preview_url = _first_present(
         source.get("preview_url"),
diff --git a/core/network_utils.py b/core/network_utils.py
index e72aa4d..ffd8d71 100644
--- a/core/network_utils.py
+++ b/core/network_utils.py
@@ -250,8 +250,8 @@ def request_source_json(
     max_attempts: int = 2,
     log_name: str = "Source API",
     raise_on_error: bool = False,
-) -> Optional[Dict[str, Any]]:
-    """Perform a GET/POST request and return JSON with retry logic for HTTP 429."""
+) -> Any:
+    """Perform a GET/POST request and return any parsed JSON payload or ``None``."""
     response = request_source_response(
         url,
         method=method,
@@ -311,10 +311,12 @@ def execute_provider_json_request(
     timeout: int = 20,
     max_attempts: int = 2,
     raise_on_error: bool = False,
-) -> Optional[Dict[str, Any]]:
+) -> Any:
     """
-    Execute an outbound JSON request to a provider (CivitAI, CivArchive, HF, etc.)
-    with normalized user-agent, authorization token/session headers, and 429 retry handling.
+    Execute an outbound JSON request and return any parsed JSON payload.
+
+    Provider APIs may return either objects or arrays. Callers that require a
+    specific shape must validate the returned value before using it.
     """
     from .type_utils import DEFAULT_BROWSER_USER_AGENT
 
diff --git a/core/sources/civitai.py b/core/sources/civitai.py
index fbe3ea6..2038a8d 100644
--- a/core/sources/civitai.py
+++ b/core/sources/civitai.py
@@ -1390,9 +1390,10 @@ def get_model_info_by_hash(
             api_url,
             api_key=api_key,
             timeout=15,
+            raise_on_error=True,
         )
 
-        if data:
+        if isinstance(data, dict) and data:
             # Extract useful info from the response
             model_info = data.get("model", {})
             version_info = data
@@ -1435,10 +1436,21 @@ def get_model_info_by_hash(
             log.info(f"Found model by hash {file_hash}: {result.get('model_name')}")
             return result
 
-        else:
-            log.info(f"Model not found or lookup failed on CivitAI for hash {file_hash}")
+        log.warning(
+            f"CivitAI hash lookup returned an unexpected JSON payload for {file_hash}"
+        )
+        return None
+
+    except requests.HTTPError as e:
+        status_code = getattr(e.response, "status_code", None)
+        if status_code == 404:
+            log.info(f"Model not found on CivitAI for hash {file_hash}")
             _hash_cache[cache_key] = None
             return None
+        log.warning(
+            f"CivitAI hash lookup returned HTTP {status_code or 'error'} for {file_hash}"
+        )
+        return None
 
     except Exception as e:
         log.error(f"Error looking up model by hash {file_hash}: {e}")
diff --git a/tests/test_downloader_metadata.py b/tests/test_downloader_metadata.py
index 36c8000..498da62 100644
--- a/tests/test_downloader_metadata.py
+++ b/tests/test_downloader_metadata.py
@@ -16,6 +16,39 @@ from core.downloader import (
 
 
 class DownloaderMetadataSidecarTests(unittest.TestCase):
+    def test_civarchive_metadata_keeps_provider_page_separate_from_hf_mirror(self):
+        with tempfile.TemporaryDirectory() as tmpdir:
+            model_path = os.path.join(tmpdir, "model.safetensors")
+            with open(model_path, "wb") as handle:
+                handle.write(b"abc")
+
+            payload = build_lora_manager_metadata(
+                model_path,
+                {
+                    "source": "civarchive",
+                    "details_source": "civarchive",
+                    "model_id": 123,
+                    "version_id": 456,
+                    "version_url": (
+                        "https://huggingface.co/author/repo/blob/main/model.safetensors"
+                    ),
+                    "url": "https://civarchive.com/models/123?modelVersionId=456",
+                    "download_url": (
+                        "https://huggingface.co/author/repo/resolve/main/model.safetensors"
+                    ),
+                    "sha256": "a" * 64,
+                },
+                category="checkpoints",
+            )
+
+        expected_page = "https://civarchive.com/models/123?modelVersionId=456"
+        self.assertEqual(expected_page, payload["source_url"])
+        self.assertEqual(expected_page, payload["version_url"])
+        self.assertEqual(
+            "https://huggingface.co/author/repo/resolve/main/model.safetensors",
+            payload["download_url"],
+        )
+
     def test_huggingface_metadata_does_not_create_civitai_identity(self):
         with tempfile.TemporaryDirectory() as tmpdir:
             model_path = os.path.join(tmpdir, "model.safetensors")
diff --git a/tests/test_downloads_tab_workflow_route.mjs b/tests/test_downloads_tab_workflow_route.mjs
index df91184..b4c4ea9 100644
--- a/tests/test_downloads_tab_workflow_route.mjs
+++ b/tests/test_downloads_tab_workflow_route.mjs
@@ -2,7 +2,8 @@ import fs from 'node:fs';
 import path from 'node:path';
 import test from 'node:test';
 import assert from 'node:assert/strict';
-import { normalizePathIdentity } from '../web/resolver/utils/html_utils.js';
+import { html, normalizePathIdentity } from '../web/resolver/utils/html_utils.js';
+import { getModelCardUrl } from '../web/resolver/utils/url_utils.js';
 
 const projectRoot = path.resolve(import.meta.dirname, '..');
 const queueMethodsSource = fs.readFileSync(
@@ -38,6 +39,15 @@ const renderFormatMethodsSource = fs.readFileSync(
   'utf8'
 );
 
+test('html template escapes every item in interpolated arrays', () => {
+  const rendered = html`<div>${['<script>', '"quoted"', '&value']}</div>`;
+
+  assert.equal(
+    rendered,
+    '<div>&lt;script&gt;&quot;quoted&quot;&amp;value</div>'
+  );
+});
+
 function extractMethod(source, methodName, paramsPattern = '[^)]*') {
   const signatureRegex = new RegExp(`\\n\\s+(async\\s+)?${methodName}\\s*\\(${paramsPattern}\\)\\s*\\{`);
   const match = signatureRegex.exec(source);
@@ -637,6 +647,89 @@ test('manual URL download metadata never inherits provider identity or hash from
   assert.equal('civitai' in metadata, false);
 });
 
+test('CivArchive metadata keeps its page URL when the download mirror is HuggingFace', () => {
+  const getDownloadMetadata = eval(`(${extractMethod(downloadTargetMethodsSource, 'getDownloadMetadata')})`);
+  const civarchivePage = 'https://civarchive.com/models/123?modelVersionId=456';
+  const huggingFaceMirror = 'https://huggingface.co/author/repo/resolve/main/model.safetensors';
+  const source = {
+    source: 'civarchive',
+    details_source: 'civarchive',
+    model_id: 123,
+    version_id: 456,
+    filename: 'model.safetensors',
+    url: civarchivePage,
+    version_url: 'https://huggingface.co/stale/repo/blob/main/model.safetensors',
+    download_url: huggingFaceMirror
+  };
+  const dialog = {
+    getCachedSearchSuggestionData() {
+      return {
+        source: 'huggingface',
+        version_url: 'https://huggingface.co/stale/repo/blob/main/model.safetensors'
+      };
+    },
+    getCompatibleCivitaiSearchResult() {
+      return {};
+    },
+    getFilenameFromPath(value = '') {
+      return String(value).split(/[\\/]/).at(-1) || '';
+    }
+  };
+
+  const metadata = getDownloadMetadata.call(dialog, {
+    original_path: source.filename,
+    category: 'checkpoints'
+  }, source, {
+    filename: source.filename,
+    category: 'checkpoints',
+    url: huggingFaceMirror,
+    openUrl: '',
+    pathMetadata: { filename: source.filename, category: 'checkpoints' }
+  });
+
+  assert.equal(metadata.source, 'civarchive');
+  assert.equal(metadata.url, civarchivePage);
+  assert.equal(metadata.version_url, civarchivePage);
+  assert.equal(metadata.page_url, civarchivePage);
+  assert.equal(metadata.download_url, huggingFaceMirror);
+});
+
+test('source link prefers the selected provider page over another provider mirror', () => {
+  const getContextMenuSourceLink = eval(`(${extractMethod(modelInfoMethodsSource, 'getContextMenuSourceLink')})`);
+  const civarchivePage = 'https://civarchive.com/models/123?modelVersionId=456';
+  const dialog = {
+    getContextMenuSourceKey() {
+      return 'civarchive';
+    },
+    getContextMenuSourceUrlCandidates() {
+      return [
+        'https://huggingface.co/author/repo/blob/main/model.safetensors',
+        civarchivePage
+      ];
+    },
+    getContextMenuSourceKeyFromUrl(value = '') {
+      if (value.includes('civarchive.com')) return 'civarchive';
+      if (value.includes('huggingface.co')) return 'huggingface';
+      return '';
+    },
+    normalizeContextMenuSourceUrl(value = '') {
+      return value;
+    },
+    getContextMenuSourceConfig(key = '') {
+      return { key, label: key === 'civarchive' ? 'CivArchive' : 'HuggingFace' };
+    },
+    buildContextMenuSourceUrlFromIds() {
+      return '';
+    }
+  };
+
+  const result = getContextMenuSourceLink.call(dialog, {});
+
+  assert.equal(result.key, 'civarchive');
+  assert.equal(result.label, 'CivArchive');
+  assert.equal(result.url, civarchivePage);
+});
+
 test('manual URL table row preserves the explicit provenance boundary', () => {
   const getCustomUrlResultTableRow = eval(`(${extractMethod(searchPanelMethodsSource, 'getCustomUrlResultTableRow')})`);
   const result = {
diff --git a/tests/test_source_internals.py b/tests/test_source_internals.py
index 8091ac1..8b7de44 100644
--- a/tests/test_source_internals.py
+++ b/tests/test_source_internals.py
@@ -16,6 +16,8 @@ import tempfile
 import unittest
 from unittest.mock import MagicMock, patch
 
+import requests
+
 # ---------------------------------------------------------------------------
 # civarchive internals
 # ---------------------------------------------------------------------------
@@ -47,6 +49,7 @@ from core.sources.civitai import (
     _search_civitai_public_api_candidates,
     _search_civitai_red_candidates,
     _search_civitai_trpc_candidates,
+    get_model_info_by_hash,
     get_model_info_for_file,
     search_civitai_for_file,
 )
@@ -95,6 +98,52 @@ class CivitaiResultBuilderTests(unittest.TestCase):
         self.assertEqual(10 * 1024, result["size"])
 
 
+class CivitaiHashLookupCacheTests(unittest.TestCase):
+
+    def setUp(self):
+        clear_search_cache()
+
+    def tearDown(self):
+        clear_search_cache()
+
+    @staticmethod
+    def _http_error(status_code):
+        response = requests.Response()
+        response.status_code = status_code
+        return requests.HTTPError(response=response)
+
+    def test_only_not_found_response_is_cached(self):
+        with patch(
+            "core.sources.civitai.execute_provider_json_request",
+            side_effect=self._http_error(404),
+        ) as request_json:
+            self.assertIsNone(get_model_info_by_hash("a" * 64))
+            self.assertIsNone(get_model_info_by_hash("a" * 64))
+
+        request_json.assert_called_once()
+        self.assertTrue(request_json.call_args.kwargs["raise_on_error"])
+
+    def test_transient_http_error_is_not_cached(self):
+        with patch(
+            "core.sources.civitai.execute_provider_json_request",
+            side_effect=self._http_error(503),
+        ) as request_json:
+            self.assertIsNone(get_model_info_by_hash("b" * 64))
+            self.assertIsNone(get_model_info_by_hash("b" * 64))
+
+        self.assertEqual(request_json.call_count, 2)
+
+    def test_unexpected_json_shape_is_not_cached(self):
+        with patch(
+            "core.sources.civitai.execute_provider_json_request",
+            return_value=[],
+        ) as request_json:
+            self.assertIsNone(get_model_info_by_hash("c" * 64))
+            self.assertIsNone(get_model_info_by_hash("c" * 64))
+
+        self.assertEqual(request_json.call_count, 2)
+
+
 # ===========================================================================
 # civitai - local safetensors fallback
 # ===========================================================================
diff --git a/tests/test_unified_helpers.py b/tests/test_unified_helpers.py
index 3e6543b..8de956f 100644
--- a/tests/test_unified_helpers.py
+++ b/tests/test_unified_helpers.py
@@ -1143,6 +1143,19 @@ class UnifiedHelpersTests(unittest.TestCase):
         res = request_source_json("https://example.com/api")
         self.assertEqual(res, {"success": True})
 
+    @patch("requests.get")
+    def test_request_source_json_accepts_array_payload(self, mock_get):
+        from core.network_utils import request_source_json
+
+        mock_response = MagicMock()
+        mock_response.status_code = 200
+        mock_response.json.return_value = [{"id": "author/model"}]
+        mock_get.return_value = mock_response
+
+        res = request_source_json("https://example.com/api")
+
+        self.assertEqual(res, [{"id": "author/model"}])
+
     @patch("requests.get")
     @patch("time.sleep")
     def test_request_source_json_429_retry(self, mock_sleep, mock_get):
@@ -1216,4 +1229,17 @@ class UnifiedHelpersTests(unittest.TestCase):
         call_kwargs = mock_request_json.call_args[1]
         self.assertIn("Cookie", call_kwargs["headers"])
 
+    @patch("core.network_utils.request_source_json")
+    def test_execute_provider_json_request_accepts_array_payload(
+        self, mock_request_json
+    ):
+        from core.network_utils import execute_provider_json_request
 
+        mock_request_json.return_value = [{"id": "author/model"}]
+
+        result = execute_provider_json_request(
+            "HuggingFace API",
+            "https://huggingface.co/api/models",
+        )
+
+        self.assertEqual(result, [{"id": "author/model"}])
diff --git a/web/resolver/search/download_target_methods.js b/web/resolver/search/download_target_methods.js
index de805c9..6f7408b 100644
--- a/web/resolver/search/download_target_methods.js
+++ b/web/resolver/search/download_target_methods.js
@@ -1412,6 +1412,39 @@ export const downloadTargetMethods = {
             || merged.details_source
             || merged.source
             || '';
+        const sourcePageCandidates = [
+            options.openUrl,
+            sourceData.version_url,
+            sourceData.versionUrl,
+            sourceData.model_url,
+            sourceData.modelUrl,
+            sourceData.page_url,
+            sourceData.pageUrl,
+            sourceData.url,
+            merged.model_url,
+            merged.page_url,
+            merged.version_url,
+            merged.url
+        ].filter(Boolean);
+        let sourcePageUrl = sourcePageCandidates[0] || '';
+        const isCivArchiveSource = String(sourceName).toLowerCase().replace(/[-\s]+/g, '_') === 'civarchive';
+        if (isCivArchiveSource) {
+            const civarchivePageUrl = sourcePageCandidates.find(value => {
+                try {
+                    const host = new URL(String(value), 'https://civarchive.com').hostname.toLowerCase();
+                    return host === 'civarchive.com' || host.endsWith('.civarchive.com');
+                } catch {
+                    return false;
+                }
+            });
+            const sourceModelId = idSource.model_id || idSource.modelId || '';
+            const sourceVersionId = idSource.version_id || idSource.versionId || '';
+            sourcePageUrl = civarchivePageUrl || (
+                sourceModelId
+                    ? `https://civarchive.com/models/${sourceModelId}${sourceVersionId ? `?modelVersionId=${sourceVersionId}` : ''}`
+                    : ''
+            );
+        }
         const metadata = {
             source: sourceName,
             details_source: sourceData.details_source || sourceData.source || sourceName,
@@ -1431,8 +1464,11 @@ export const downloadTargetMethods = {
             creator,
             author: merged.author || pathMetadata.author || '',
             repo_id: repoId,
-            url: options.openUrl || merged.model_url || merged.version_url || merged.url || '',
-            version_url: merged.version_url || '',
+            url: sourcePageUrl,
+            version_url: isCivArchiveSource ? sourcePageUrl : (sourceData.version_url || sourceData.versionUrl || sourcePageUrl),
+            page_url: isCivArchiveSource ? sourcePageUrl : (sourceData.page_url || sourceData.pageUrl || sourcePageUrl),
+            model_url: isCivArchiveSource ? sourcePageUrl : (sourceData.model_url || sourceData.modelUrl || sourcePageUrl),
+            platform_url: sourceData.platform_url || sourceData.platformUrl || merged.platform_url || merged.platformUrl || '',
             download_url: options.url || merged.download_url || merged.downloadUrl || merged.url || '',
             size: merged.size || merged.size_bytes || '',
             sha256: merged.sha256 || merged.hash || merged.hashes?.SHA256 || merged.hashes?.sha256 || '',
diff --git a/web/resolver/utils/html_utils.js b/web/resolver/utils/html_utils.js
index 0ab08b6..c59b238 100644
--- a/web/resolver/utils/html_utils.js
+++ b/web/resolver/utils/html_utils.js
@@ -27,7 +27,9 @@ export function html(strings, ...values) {
         const value = values[i - 1];
         let valStr = '';
         if (value !== undefined && value !== null) {
-            valStr = Array.isArray(value) ? value.join('') : escapeHtml(value);
+            valStr = Array.isArray(value)
+                ? value.map(item => escapeHtml(item)).join('')
+                : escapeHtml(value);
         }
         return result + valStr + string;
     });
diff --git a/web/resolver/views/model_info_methods.js b/web/resolver/views/model_info_methods.js
index feae9b8..9fce1f9 100644
--- a/web/resolver/views/model_info_methods.js
+++ b/web/resolver/views/model_info_methods.js
@@ -291,11 +291,10 @@ export const modelInfoMethods = {
     getContextMenuSourceLink(model = {}) {
         const fallbackKey = this.getContextMenuSourceKey(model);
         const candidates = this.getContextMenuSourceUrlCandidates(model);
-
-        for (const candidate of candidates) {
+        const resolveCandidate = (candidate) => {
             const candidateKey = this.getContextMenuSourceKeyFromUrl(candidate) || fallbackKey;
             const normalizedUrl = this.normalizeContextMenuSourceUrl(candidate, candidateKey);
-            if (!normalizedUrl) continue;
+            if (!normalizedUrl) return null;
 
             const cardUrl = getModelCardUrl(normalizedUrl) || normalizedUrl;
             const sourceKey = this.getContextMenuSourceKeyFromUrl(cardUrl)
@@ -306,6 +305,17 @@ export const modelInfoMethods = {
                 ...config,
                 url: cardUrl
             };
+        };
+
+        const preferredCandidates = fallbackKey
+            ? candidates.filter(candidate => {
+                const candidateKey = this.getContextMenuSourceKeyFromUrl(candidate);
+                return !candidateKey || candidateKey === fallbackKey;
+            })
+            : candidates;
+        for (const candidate of preferredCandidates) {
+            const resolved = resolveCandidate(candidate);
+            if (resolved) return resolved;
         }
 
         const builtUrl = this.buildContextMenuSourceUrlFromIds(model, fallbackKey);
@@ -317,6 +327,12 @@ export const modelInfoMethods = {
             };
         }
 
+        for (const candidate of candidates) {
+            if (preferredCandidates.includes(candidate)) continue;
+            const resolved = resolveCandidate(candidate);
+            if (resolved) return resolved;
+        }
+
         return null;
     },
 
@@ -2096,11 +2112,11 @@ export const modelInfoMethods = {
         const civitaiLinkEl = dialog.querySelector('.mr-info-civitai-link');
         if (civitaiLinkEl) {
             const hasLocalFile = Boolean(this.getInfoDialogModelFilePath(data));
-            if (data.url || data.version_url) {
-                const url = data.version_url || data.url;
-                const sourceLabel = this.getInfoMetadataSourceLabel(data);
+            const sourceLink = this.getContextMenuSourceLink(data);
+            if (sourceLink?.url) {
+                const sourceLabel = sourceLink.label || this.getInfoMetadataSourceLabel(data);
                 civitaiLinkEl.innerHTML = `
-                    <a href="${this.escapeHtml(url)}" target="_blank" class="mr-info-link">
+                    <a href="${this.escapeHtml(sourceLink.url)}" target="_blank" class="mr-info-link">
                         View on ${this.escapeHtml(sourceLabel)}
                         ${this.getInfoMetadataSourceIcon(data)}
                     </a>