Merge pull request #26 from anurag162008/codex/investigate-syncing-issue-with-huggingclaw-backup

.env.example

@@ -175,6 +175,13 @@ LLM_API_KEY_FALLBACK_ENABLED=true
 #      is also picked up by OpenClaw → that provider's models become
 #      available in the Control UI for manual selection.
 #   3. Rotation pools (*_API_KEYS) work for every active provider
+#
+# Optional: explicitly pin model lists per provider for Control UI visibility
+# when provider keys are configured.
+# Format: comma-separated model IDs
+# NVIDIA_MODELS=meta/llama-3.1-70b-instruct,nvidia/llama-3.1-nemotron-70b-instruct
+# OPENAI_MODELS=gpt-4o-mini,gpt-4.1
+# GROQ_MODELS=llama-3.3-70b-versatile,deepseek-r1-distill-llama-70b
 #      independently and in parallel.
 #
 # EXAMPLE — default Anthropic, also use OpenAI and Groq:

ENV_VARIABLES_FULL_LIST.md

@@ -75,6 +75,9 @@ Full list:
 - `XIAOMI_API_KEYS`
 - `VOLCANO_ENGINE_API_KEYS`
 - `BYTEPLUS_API_KEYS`
+- `QIANFAN_API_KEYS`
+- `MODELSTUDIO_API_KEYS`
+- `KIMI_API_KEYS`
 - `MISTRAL_API_KEYS`
 - `XAI_API_KEYS`
 - `NVIDIA_API_KEYS`
@@ -84,6 +87,38 @@ Full list:
 - `CEREBRAS_API_KEYS`
 - `HUGGINGFACE_HUB_TOKENS`
 
+### Provider model lists (optional, for Control UI visibility)
+Format: comma-separated model IDs; used only when that provider has key(s) configured.
+Examples:
+- `NVIDIA_MODELS=meta/llama-3.1-70b-instruct,nvidia/llama-3.1-nemotron-70b-instruct`
+- `OPENAI_MODELS=gpt-4o-mini,gpt-4.1`
+- `GROQ_MODELS=llama-3.3-70b-versatile,deepseek-r1-distill-llama-70b`
+
+Supported:
+- `ANTHROPIC_MODELS`
+- `OPENAI_MODELS`
+- `GEMINI_MODELS`
+- `DEEPSEEK_MODELS`
+- `OPENROUTER_MODELS`
+- `KILOCODE_MODELS`
+- `OPENCODE_MODELS`
+- `ZAI_MODELS`
+- `MOONSHOT_MODELS`
+- `KIMI_MODELS`
+- `MINIMAX_MODELS`
+- `MODELSTUDIO_MODELS`
+- `XIAOMI_MODELS`
+- `VOLCANO_ENGINE_MODELS`
+- `BYTEPLUS_MODELS`
+- `QIANFAN_MODELS`
+- `GROQ_MODELS`
+- `MISTRAL_MODELS`
+- `XAI_MODELS`
+- `NVIDIA_MODELS`
+- `COHERE_MODELS`
+- `TOGETHER_MODELS`
+- `CEREBRAS_MODELS`
+
 ### Telegram / WhatsApp
 - `TELEGRAM_BOT_TOKEN` — bot token string. Example: `TELEGRAM_BOT_TOKEN=123456:ABC...`
 - `TELEGRAM_ALLOWED_USERS` — comma-separated numeric IDs. Example: `TELEGRAM_ALLOWED_USERS=12345,67890`

openclaw-sync.py

@@ -360,6 +360,30 @@ def create_snapshot_dir(source_root: Path) -> Path:
     return staging_root
 
 
+def prune_remote_deleted_files(repo_id: str, snapshot_dir: Path) -> None:
+    if HF_API is None:
+        return
+
+    local_files = {
+        path.relative_to(snapshot_dir).as_posix()
+        for path in snapshot_dir.rglob("*")
+        if path.is_file()
+    }
+
+    remote_files = HF_API.list_repo_files(repo_id=repo_id, repo_type="dataset")
+    stale_files = [
+        path for path in remote_files
+        if path not in local_files and path not in {".gitattributes"}
+    ]
+    for stale_path in stale_files:
+        HF_API.delete_file(
+            path_in_repo=stale_path,
+            repo_id=repo_id,
+            repo_type="dataset",
+            commit_message=f"Delete stale file {stale_path}",
+        )
+
+
 def restore_workspace() -> bool:
     if not HF_TOKEN:
         write_status("disabled", "HF_TOKEN is not configured.")
@@ -459,6 +483,7 @@ def _sync_once_unlocked(
                 commit_message=f"HuggingClaw sync {time.strftime('%Y-%m-%dT%H:%M:%SZ', time.gmtime())}",
                 ignore_patterns=[".git/*", ".git"],
             )
+        prune_remote_deleted_files(repo_id, snapshot_dir)
     finally:
         shutil.rmtree(snapshot_dir, ignore_errors=True)
 

start.sh

@@ -131,6 +131,48 @@ case "$LLM_PROVIDER" in
     ;;
 esac
 
+# Ensure OpenClaw provider discovery can see per-provider keys even when users
+# configure only *_API_KEYS pools. Mirror first pool key into singular env.
+promote_first_pool_key() {
+  local singular_var="$1"
+  local pool_var="$2"
+  local singular_val="${!singular_var:-}"
+  local pool_val="${!pool_var:-}"
+
+  [ -n "$singular_val" ] && return 0
+  [ -n "$pool_val" ] || return 0
+
+  local first
+  first=$(printf '%s' "$pool_val" | tr ',' '\n' | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//' | awk 'NF{print; exit}')
+  [ -n "$first" ] || return 0
+  export "${singular_var}=$first"
+}
+
+promote_first_pool_key "ANTHROPIC_API_KEY" "ANTHROPIC_API_KEYS"
+promote_first_pool_key "OPENAI_API_KEY" "OPENAI_API_KEYS"
+promote_first_pool_key "GEMINI_API_KEY" "GEMINI_API_KEYS"
+promote_first_pool_key "DEEPSEEK_API_KEY" "DEEPSEEK_API_KEYS"
+promote_first_pool_key "OPENROUTER_API_KEY" "OPENROUTER_API_KEYS"
+promote_first_pool_key "KILOCODE_API_KEY" "KILOCODE_API_KEYS"
+promote_first_pool_key "OPENCODE_API_KEY" "OPENCODE_API_KEYS"
+promote_first_pool_key "ZAI_API_KEY" "ZAI_API_KEYS"
+promote_first_pool_key "MOONSHOT_API_KEY" "MOONSHOT_API_KEYS"
+promote_first_pool_key "MINIMAX_API_KEY" "MINIMAX_API_KEYS"
+promote_first_pool_key "XIAOMI_API_KEY" "XIAOMI_API_KEYS"
+promote_first_pool_key "VOLCANO_ENGINE_API_KEY" "VOLCANO_ENGINE_API_KEYS"
+promote_first_pool_key "BYTEPLUS_API_KEY" "BYTEPLUS_API_KEYS"
+promote_first_pool_key "QIANFAN_API_KEY" "QIANFAN_API_KEYS"
+promote_first_pool_key "MODELSTUDIO_API_KEY" "MODELSTUDIO_API_KEYS"
+promote_first_pool_key "KIMI_API_KEY" "KIMI_API_KEYS"
+promote_first_pool_key "MISTRAL_API_KEY" "MISTRAL_API_KEYS"
+promote_first_pool_key "XAI_API_KEY" "XAI_API_KEYS"
+promote_first_pool_key "NVIDIA_API_KEY" "NVIDIA_API_KEYS"
+promote_first_pool_key "GROQ_API_KEY" "GROQ_API_KEYS"
+promote_first_pool_key "COHERE_API_KEY" "COHERE_API_KEYS"
+promote_first_pool_key "TOGETHER_API_KEY" "TOGETHER_API_KEYS"
+promote_first_pool_key "CEREBRAS_API_KEY" "CEREBRAS_API_KEYS"
+promote_first_pool_key "HUGGINGFACE_HUB_TOKEN" "HUGGINGFACE_HUB_TOKENS"
+
 # ── Setup directories ──
 mkdir -p /home/node/.openclaw/agents/main/sessions
 mkdir -p /home/node/.openclaw/credentials
@@ -285,6 +327,67 @@ if [ -n "$CUSTOM_PROVIDER_NAME" ] || [ -n "$CUSTOM_BASE_URL" ] || [ -n "$CUSTOM_
   fi
 fi
 
+# Optional: explicitly expose provider model lists in Control UI when
+# provider keys are configured. Format:
+#   NVIDIA_MODELS=model1,model2
+#   OPENAI_MODELS=gpt-4o-mini,gpt-4.1
+# This helps when provider auto-discovery does not populate models reliably.
+inject_provider_models_from_env() {
+  local provider="$1"
+  local models_env="$2"
+  local key_env_single="$3"
+  local key_env_pool="$4"
+  local models_csv="${!models_env:-}"
+  local single_key="${!key_env_single:-}"
+  local pool_keys="${!key_env_pool:-}"
+
+  # Only inject when both:
+  # 1) provider has at least one configured key
+  # 2) explicit model list env is provided
+  if [ -z "$models_csv" ] || { [ -z "$single_key" ] && [ -z "$pool_keys" ]; }; then
+    return 0
+  fi
+
+  local models_json
+  models_json=$(printf '%s' "$models_csv" \
+    | tr ',' '\n' \
+    | sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//' \
+    | awk 'NF' \
+    | jq -R . \
+    | jq -s 'map({id: ., name: .}) | unique_by(.id)')
+
+  CONFIG_JSON=$(jq \
+    --arg provider "$provider" \
+    --argjson models "$models_json" \
+    '.models.mode = "merge"
+     | .models.providers[$provider] = ((.models.providers[$provider] // {}) + {models: $models})' <<<"$CONFIG_JSON")
+}
+
+# Built-in provider model envs (optional)
+inject_provider_models_from_env "anthropic" "ANTHROPIC_MODELS" "ANTHROPIC_API_KEY" "ANTHROPIC_API_KEYS"
+inject_provider_models_from_env "openai" "OPENAI_MODELS" "OPENAI_API_KEY" "OPENAI_API_KEYS"
+inject_provider_models_from_env "google" "GEMINI_MODELS" "GEMINI_API_KEY" "GEMINI_API_KEYS"
+inject_provider_models_from_env "deepseek" "DEEPSEEK_MODELS" "DEEPSEEK_API_KEY" "DEEPSEEK_API_KEYS"
+inject_provider_models_from_env "openrouter" "OPENROUTER_MODELS" "OPENROUTER_API_KEY" "OPENROUTER_API_KEYS"
+inject_provider_models_from_env "kilocode" "KILOCODE_MODELS" "KILOCODE_API_KEY" "KILOCODE_API_KEYS"
+inject_provider_models_from_env "opencode" "OPENCODE_MODELS" "OPENCODE_API_KEY" "OPENCODE_API_KEYS"
+inject_provider_models_from_env "zai" "ZAI_MODELS" "ZAI_API_KEY" "ZAI_API_KEYS"
+inject_provider_models_from_env "moonshot" "MOONSHOT_MODELS" "MOONSHOT_API_KEY" "MOONSHOT_API_KEYS"
+inject_provider_models_from_env "kimi-coding" "KIMI_MODELS" "KIMI_API_KEY" "KIMI_API_KEYS"
+inject_provider_models_from_env "minimax" "MINIMAX_MODELS" "MINIMAX_API_KEY" "MINIMAX_API_KEYS"
+inject_provider_models_from_env "modelstudio" "MODELSTUDIO_MODELS" "MODELSTUDIO_API_KEY" "MODELSTUDIO_API_KEYS"
+inject_provider_models_from_env "xiaomi" "XIAOMI_MODELS" "XIAOMI_API_KEY" "XIAOMI_API_KEYS"
+inject_provider_models_from_env "volcengine" "VOLCANO_ENGINE_MODELS" "VOLCANO_ENGINE_API_KEY" "VOLCANO_ENGINE_API_KEYS"
+inject_provider_models_from_env "byteplus" "BYTEPLUS_MODELS" "BYTEPLUS_API_KEY" "BYTEPLUS_API_KEYS"
+inject_provider_models_from_env "qianfan" "QIANFAN_MODELS" "QIANFAN_API_KEY" "QIANFAN_API_KEYS"
+inject_provider_models_from_env "groq" "GROQ_MODELS" "GROQ_API_KEY" "GROQ_API_KEYS"
+inject_provider_models_from_env "mistral" "MISTRAL_MODELS" "MISTRAL_API_KEY" "MISTRAL_API_KEYS"
+inject_provider_models_from_env "xai" "XAI_MODELS" "XAI_API_KEY" "XAI_API_KEYS"
+inject_provider_models_from_env "nvidia" "NVIDIA_MODELS" "NVIDIA_API_KEY" "NVIDIA_API_KEYS"
+inject_provider_models_from_env "cohere" "COHERE_MODELS" "COHERE_API_KEY" "COHERE_API_KEYS"
+inject_provider_models_from_env "together" "TOGETHER_MODELS" "TOGETHER_API_KEY" "TOGETHER_API_KEYS"
+inject_provider_models_from_env "cerebras" "CEREBRAS_MODELS" "CEREBRAS_API_KEY" "CEREBRAS_API_KEYS"
+
 # Browser configuration (managed local Chromium in HF/Docker)
 BROWSER_EXECUTABLE_PATH=""
 for candidate in /usr/bin/chromium /usr/bin/chromium-browser /snap/bin/chromium; do