fix fresh-host acceptance and document real-host debugging learnings

scripts/import_remote43_provider.sh

@@ -41,6 +41,20 @@ if [[ -z "$upstream_key" ]]; then
   exit 2
 fi
 
+upstream_base_url="$(python3 - "$PACK_PATH" "$provider_id" <<'PY'
+import json, pathlib, sys
+pack_path = pathlib.Path(sys.argv[1])
+provider_id = sys.argv[2]
+provider_file = pack_path / "providers" / f"{provider_id}.json"
+provider = json.loads(provider_file.read_text(encoding='utf-8'))
+print(str(provider.get("base_url", "")).strip())
+PY
+)"
+if [[ -z "$upstream_base_url" ]]; then
+  echo "missing provider base_url for $provider_id in $PACK_PATH" >&2
+  exit 2
+fi
+
 ssh_cmd() {
   local cmd="$1"
   ssh -i "$KEY" -o StrictHostKeyChecking=no "$REMOTE" "$cmd"
@@ -454,6 +468,14 @@ ssh_cmd "curl -sS -D /tmp/chat_headers.txt -o /tmp/chat_body.json -H 'Authorizat
 ssh_cmd "cat /tmp/chat_headers.txt" > "$ART/11-chat.headers.txt"
 ssh_cmd "cat /tmp/chat_body.json" > "$ART/12-chat.body.json"
 
+ssh_cmd "curl -sS -D /tmp/upstream_models_headers.txt -o /tmp/upstream_models_body.json -H 'Authorization: Bearer $upstream_key' ${upstream_base_url%/}/models"
+ssh_cmd "cat /tmp/upstream_models_headers.txt" > "$ART/17-upstream-models.headers.txt"
+ssh_cmd "cat /tmp/upstream_models_body.json" > "$ART/18-upstream-models.body.json"
+
+ssh_cmd "curl -sS -D /tmp/upstream_chat_headers.txt -o /tmp/upstream_chat_body.txt -H 'Authorization: Bearer $upstream_key' -H 'Content-Type: application/json' ${upstream_base_url%/}/chat/completions -d $(printf %q "$probe_payload")"
+ssh_cmd "cat /tmp/upstream_chat_headers.txt" > "$ART/19-upstream-chat.headers.txt"
+ssh_cmd "cat /tmp/upstream_chat_body.txt" > "$ART/20-upstream-chat.body.txt"
+
 crm_curl_json GET "/api/providers/$provider_id/status" > "$ART/13-provider-status.json"
 crm_curl_json GET "/api/providers/$provider_id/access/status" > "$ART/14-access-status.json"
 preview_payload="$(python3 - "$provider_id" <<'PY'
@@ -472,17 +494,67 @@ provider_id=sys.argv[2]
 batch_id=int(sys.argv[3])
 subscription_group_id=sys.argv[4]
 expected_model=sys.argv[5]
+
+def normalize_model_id(model_id: str) -> str:
+    value = str(model_id or '').strip().lower()
+    if not value:
+        return ''
+    if '/' in value:
+        value = value.split('/')[-1]
+    return value
+
+def has_expected_model(models, expected: str) -> bool:
+    normalized_expected = normalize_model_id(expected)
+    if not normalized_expected:
+        return False
+    return any(normalize_model_id(model_id) == normalized_expected for model_id in models)
+
+def status_from_headers(path: pathlib.Path) -> int:
+    if not path.exists():
+        return 0
+    for line in path.read_text(encoding='utf-8').splitlines():
+        parts = line.strip().split()
+        if len(parts) >= 2 and parts[0].startswith('HTTP/'):
+            try:
+                return int(parts[1])
+            except ValueError:
+                return 0
+    return 0
+
+def load_json(path: pathlib.Path):
+    try:
+        return json.loads(path.read_text(encoding='utf-8'))
+    except Exception:
+        return {}
+
 import_obj=json.loads((art/'03-import.body.json').read_text())
-models_obj=json.loads((art/'10-models.body.json').read_text())
-access_status=json.loads((art/'14-access-status.json').read_text())
-preview=json.loads((art/'15-access-preview.json').read_text())
+models_obj=load_json(art/'10-models.body.json')
+access_status=load_json(art/'14-access-status.json')
+preview=load_json(art/'15-access-preview.json')
 models_headers=(art/'09-models.headers.txt').read_text()
 chat_headers=(art/'11-chat.headers.txt').read_text()
+upstream_models_obj=load_json(art/'18-upstream-models.body.json')
+upstream_chat_headers=(art/'19-upstream-chat.headers.txt')
+upstream_chat_body=(art/'20-upstream-chat.body.txt').read_text(encoding='utf-8')
 models=[]
 for item in models_obj.get('data') or []:
     model_id = item.get('id')
     if isinstance(model_id, str) and model_id:
         models.append(model_id)
+upstream_models=[]
+for item in upstream_models_obj.get('data') or []:
+    model_id = item.get('id')
+    if isinstance(model_id, str) and model_id:
+        upstream_models.append(model_id)
+host_chat_status = status_from_headers(art/'11-chat.headers.txt')
+upstream_chat_status = status_from_headers(upstream_chat_headers)
+classification = 'unknown'
+direct_has_expected_model = has_expected_model(models, expected_model)
+upstream_has_expected_model = has_expected_model(upstream_models, expected_model)
+if direct_has_expected_model and host_chat_status >= 500 and upstream_chat_status == 200:
+    classification = 'host_compatibility_gap'
+elif direct_has_expected_model and upstream_chat_status == 403 and 'insufficient_user_quota' in upstream_chat_body:
+    classification = 'upstream_key_quota_issue'
 summary={
   'artifact_dir': str(art),
   'provider_id': provider_id,
@@ -491,14 +563,21 @@ summary={
   'access_status_from_import': import_obj.get('access_status'),
   'provider_status_from_import': import_obj.get('provider_status'),
   'direct_models_http200': '200 OK' in models_headers,
-  'direct_models_has_expected_model': expected_model in models,
+  'direct_models_has_expected_model': direct_has_expected_model,
   'direct_models': models,
   'direct_chat_http200': '200 OK' in chat_headers,
+  'direct_chat_status': host_chat_status,
+  'upstream_models': upstream_models,
+  'upstream_models_has_expected_model': upstream_has_expected_model,
+  'upstream_chat_status': upstream_chat_status,
+  'completion_classification': classification,
   'latest_access_status': access_status.get('latest_access_status') or access_status.get('batch_access_status'),
   'preview_available': preview.get('available'),
   'accepted_keys_count': import_obj.get('accepted_keys_count'),
   'subscription_group_id': subscription_group_id,
   'import_group_id': (import_obj.get('group') or {}).get('id'),
 }
-print(json.dumps(summary, ensure_ascii=False))
+summary_json = json.dumps(summary, ensure_ascii=False)
+(art / '21-summary.json').write_text(summary_json, encoding='utf-8')
+print(summary_json)
 PY