@clickzetta/cz-cli-darwin-x64 0.5.16 → 0.5.18

package/bin/skills/lakehouse-doc-en/references/aigw-2026-2-5-1.0.md

@@ -0,0 +1,16 @@
+## V1.0 Release Notes
+
+**Release Date**: February 5, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+## Core Features Launch
+
+**Release Scope**:
+
+1. Standardized unified API gateway: Full compatibility with the OpenAI interface protocol. Business systems built on OpenAI can integrate without any code modifications.
+2. First-wave mainstream model onboarding: Integrated text large language models including ByteDance Doubao, Alibaba Cloud Qwen, and DeepSeek.
+3. Basic key management: Supports full lifecycle management of model API keys on a single platform, including binding, enabling, and disabling.
+4. Basic routing and scheduling: Supports manual specification of model call providers and load balancing across multiple providers.
+5. Basic monitoring and statistics: Provides real-time visualization of core metrics such as call volume and call success rate.
+
+***

package/bin/skills/lakehouse-doc-en/references/aigw-2026-3-29-1.0.2.md

@@ -0,0 +1,14 @@
+## V1.0.2 Release Notes
+
+**Release Date**: March 29, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+## Core Features Launch
+
+**Release Scope**:
+
+1. Fully standardized interfaces: Unified request and response formats across all integrated models, enabling cross-model calls with no interface differences.
+2. Billing management: Supports generating billing statements by model and by time dimension, with billing export functionality.
+3. Call log traceability: Supports querying, filtering, and exporting all call logs within the past 7 days.
+4. Multi-language development support: Released SDKs and complete call examples for mainstream languages including Python, Java, and Go.
+5. Online help center: Launched the product user manual and FAQ library.

package/bin/skills/lakehouse-doc-en/references/aigw-2026-3-8-1.0.1.md

@@ -0,0 +1,16 @@
+## V1.0.1 Release Notes
+
+**Release Date**: March 8, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+## Core Features Launch
+
+**Release Scope**:
+
+1. Unified multi-model key management: A single API key can be bound to multiple platform models simultaneously, enabling one key to manage all large model call permissions.
+2. BYOK (Bring Your Own Key) support: Integrate your own model API keys. Models can be called through the unified platform without migration.
+3. Cost-optimized smart routing: New price-based routing strategy that automatically selects the lowest-cost available provider.
+4. First model library expansion: Added high cost-efficiency text models including Qwen 3.5 and DeepSeek V2; added image generation models including Tongyi Wanxiang 2.0, officially enabling text-to-image capabilities.
+5. Basic permission isolation: Supports restricting the callable model list per API key, enabling permission isolation across different business lines.
+
+***

package/bin/skills/lakehouse-doc-en/references/aigw-2026-4-28-1.1.md

@@ -0,0 +1,29 @@
+## V1.1 Release Notes
+
+**Release Date**: April 28, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+## Core Features Launch
+
+1. **Unified model calling platform**: A single standardized API can call all large model services provided by the platform, supporting multi-modal capabilities including text generation, image generation, and video generation.
+
+2. **Flexible API Key binding**
+
+   * A single API Key can be bound to multiple platform models simultaneously, eliminating the need to create separate keys for different models.
+   * Supports binding user-owned models (BYOK). Models can be called through the unified platform without migration, offering greater flexibility.
+
+3. **Smart routing strategy configuration**
+
+   * Price-based routing strategy: automatically selects the lowest-cost available provider to reduce call costs.
+   * Provider-based routing strategy: supports customizing provider priority so the same model is automatically dispatched to the specified provider.
+   * Custom model visibility: restricts the list of callable models per API Key.
+
+4. **Core management capabilities**
+
+   * Visual usage statistics and billing query.
+   * Endpoint start/stop control for flexible business access management.
+   * Complete online interface documentation and call examples.
+
+## Upcoming Features
+
+We will continue to optimize the platform experience. Upcoming features include a revamped model marketplace, enhanced batch inference capabilities, and expanded vector models. Stay tuned.

package/bin/skills/lakehouse-doc-en/references/aigw-2026-5-12-1.1.1.md

@@ -0,0 +1,18 @@
+## V1.1.1 Release Notes
+
+**Release Date**: May 12, 2026
+**Release Scope**: Simultaneous release on Alibaba Cloud Shanghai and AWS Singapore
+
+## Core Upgrades
+
+This release completes a **full underlying architecture overhaul**, reinforcing stability at the service layer and significantly improving request success rates and service reliability in high-concurrency scenarios, providing a more stable model invocation experience.
+
+## Feature Updates
+
+1. **New BYOK (Bring Your Own Key) visual management interface** — Users can now upload their own API keys directly in the platform console and enable connectivity with platform-integrated models in one click, without requiring backend configuration. The process is simpler and more intuitive.
+2. **API key quota supports long-term validity** — Added a long-term validity option for API key quotas, eliminating the need for periodic renewals to support stable long-term use.
+3. **Pass-Through mode** — Supports native protocol pass-through for third-party models, enabling compatibility with more third-party service call methods.
+
+### Important Notes
+
+1. To ensure service stability and reliability, V1.1 and later will no longer support user-defined addition of models and providers that have not been officially integrated through the platform.

package/bin/skills/lakehouse-doc-en/references/aigw-2026-5-15-1.2.md

@@ -0,0 +1,9 @@
+## V1.2 Release Notes
+
+**Release Date**: May 15, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+### Feature Updates
+
+1. **Smart routing strategy configuration** supports configuring routing rules by provider priority. The same model can be automatically dispatched to specified provider resources.
+2. **Online API documentation** releases complete feature documentation and API call examples.

package/bin/skills/lakehouse-doc-en/references/aigw-2026-5-21-1.3.md

@@ -0,0 +1,9 @@
+## V1.3 Release Notes
+
+**Release Date**: May 21, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+### Feature Updates
+
+1. **Platform model library expansion**: Added the Seedream 5.0 Lite image generation model, supporting text-to-image and image-to-image capabilities.
+2. **Improved interface error messages**: Error responses now include request retry history, making it easier to quickly identify the cause of failed calls.

package/bin/skills/lakehouse-doc-en/references/aigw-2026-5-28-1.4.md

@@ -0,0 +1,10 @@
+## V1.4 Release Notes
+
+**Release Date**: May 28, 2026
+**Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+### Feature Updates
+
+1. **Endpoint start/stop control**: Supports independently enabling or disabling the corresponding call endpoint addresses, enabling flexible control over business activation and deactivation.
+2. **Model filtering**: Supports quickly filtering target models by model type and provider.
+3. **Improved operation experience**: The key management page now includes a one-click API key copy feature, simplifying the configuration process.

package/bin/skills/lakehouse-doc-en/references/aigw-2026-6-3-1.5.md

@@ -0,0 +1,9 @@
+## V1.5 Release Notes
+
+**Release Date**: June 3, 2026  **Release Scope**: Simultaneous release across Alibaba Cloud Shanghai and AWS Singapore regions
+
+### Feature Updates
+
+1. **Video generation officially launched**: Supports asynchronous video generation task submission, progress querying, result retrieval, and deletion.
+2. **Whitelist system launched**: Implements system-level model whitelist and API key whitelist management for fine-grained access control.
+3. **Platform model library continues to expand**: Added multiple new video and image generation models to enrich customer selection options.

package/bin/skills/lakehouse-doc-en/references/alicloud-arn-externalid.md

@@ -8,7 +8,6 @@ When configuring Private Link to access the Lakehouse network, to ensure that th
 
 First, go to the Alibaba Cloud Access Control page (<https://ram.console.aliyun.com/roles>) and click "Create Role", then select the feasible entity type as: Alibaba Cloud Account:
 
-:-: ![](.topwrite/assets/image_1733068362672.png =424)
 
 ^
 
@@ -16,25 +15,21 @@ Enter a custom role name in the role name field;
 
 In "Select Trusted Cloud Account", choose "Other Cloud Account", and copy the Lakehouse endpoint service's LakehouseUID content and paste it in.
 
-:-: ![](.topwrite/assets/image_1733068490413.png =769)
 
 ^
 
 After the role is created, click "Authorize Role", and then click "Create Authorization";
 
-:-: ![](.topwrite/assets/image_1733068711869.png =799)
 
 ^
 
 Search for "privatelink", and check: AliyunPrivateLinkReadOnlyAccess and AliyunPrivatelinkEndpointServiceReadOnlyAccess two policies;
 
-:-: ![](.topwrite/assets/image_1733068776650.png =811)
 
 ^
 
 After completing "Confirm New Authorization", switch to the "Trust Policy" tab, and click "Edit Trust Policy":
 
-:-: ![](.topwrite/assets/image_1733068938505.png =812)
 
 Paste the following content between "Action": "sts:AssumeRole", and "Effect": "Allow":
 Sure, here is the translated content:

package/bin/skills/lakehouse-doc-en/references/answer-accuracy-improve.md

@@ -1,179 +1,196 @@
-# Analytics Agent Answer Accuracy Improvement - Best Practices
+# Analytics Agent Answer Accuracy Improvement Best Practices
 
-In response to the most frequently asked question: What methods are available to improve answer accuracy? Which scenarios are suitable for each method? How to manage them?
+Addressing the most common question: what techniques can improve answer accuracy, which approach fits which scenario, and how to manage them?
+
+## 0. Reduce Noise at the Architecture Level: Partition Analysis Domains
+
+The single most effective first step to improving accuracy is not tuning parameters — it is narrowing the model's "search space." When tables from all business lines are piled together, the model has to guess the right table from dozens or even hundreds of candidates every time it answers — the more noise, the higher the chance of selecting the wrong one.
+
+Analysis domains let you split data by business topic: one domain for sales, one for HR, one for supply chain. Each domain contains only its relevant tables, metrics, and knowledge. When a user asks a question within a domain, the model reasons only within that domain's scope, which naturally improves accuracy. Domains can also be bound to users, providing built-in data isolation — users from different departments can only see data within their own domain.
+
+:-: ![](/.topwrite/assets/image_1780899840394.png =697)
+
+The **Analysis** page shows all analysis domains, each displaying the count of tables, metrics, answer builders, and knowledge items it contains. Click **New Analysis Domain** to create one, then associate data tables and users in the domain settings.
+
+> 💡 **Tip**: If you have only one business scenario, a single domain is sufficient. Consider splitting only when the model frequently selects the wrong table, or when you need to isolate data for different teams.
+
+***
 
 ## 1. What Capabilities Improve Accuracy?
 
-Analytics Agent (formerly known as DataGPT) provides **4 core capabilities** to give the model more business information input and improve answer accuracy:
+Data Analytics Agent (DataGPT) provides **4 core capabilities** to feed more business context to the model and improve answer accuracy:
 
-### 1. Table/Column Schema Annotations
+### 1. Table/Column Schema Descriptions
 
-![](/.topwrite/assets/image_1779160426141.png)
+:-: ![](/.topwrite/assets/image_1780900044863.png =738)
 
-| Configuration Item              | Purpose                                                                         | Example                                           |
-| ------------------------------- | ------------------------------------------------------------------------------- | ------------------------------------------------- |
-| Table Description               | Tells the model what this table represents                                      | "Daily sales transaction records for all stores in Hong Kong region" |
-| Column Alias                    | Helps the model understand different names for the same field                   | Column name rev\_amt → alias "revenue", "income"            |
-| Column Description              | Explains the business meaning and value rules of a column                       | "payment\_type: payment method, values include cash/card/octopus" |
-| Hidden Column                   | When a table has many fields, some columns are irrelevant for business analysis (e.g., system primary keys, ETL timestamps, internal codes). Hiding them prevents them from participating in AI Q&A and analysis, reducing the chance of the model selecting the wrong column | Applicable scenarios:
Tables with many technical fields that interfere with column selection; or sensitive fields that should not be referenced by AI |
-| Virtual Column                  | When schema changes require AI to adapt its understanding, create virtual columns through SQL expressions without modifying the underlying data table | Use case: When raw fields need transformation to help AI understand business meaning |
+| Configuration      | Purpose                                                                                                                                                                                                                                                       | Example                                                                                                                          |
+| ------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
+| Table description  | Tell the model what this table contains                                                                                                                                                                                                                       | "Daily sales transaction data for all stores in the Hong Kong region"                                                            |
+| Column alias       | Help the model understand different names for the same field                                                                                                                                                                                                  | Column `rev_amt` → alias "revenue", "turnover"                                                                                   |
+| Column description | Explain business meaning and value rules for a column                                                                                                                                                                                                         | "`payment_type`: payment method, values include cash/card/octopus"                                                               |
+| Hidden columns     | When a table has many fields, some columns are meaningless for business analysis (e.g., system primary keys, ETL timestamps, internal codes). Hiding them prevents these columns from participating in AI Q\&A, reducing the chance of wrong column selection | Use when: the table has many technical fields that confuse the model; or certain sensitive fields should not be referenced by AI |
+| Virtual columns    | When schema changes require AI to adapt its understanding, create a virtual column via a SQL expression without modifying the underlying table                                                                                                                | Use when: a raw field needs transformation to make its business meaning more understandable to AI                                |
 
 ### 2. Knowledge (Knowledge Base / Analysis Documents)
 
-Business background knowledge provided in file form, injected into conversation context via RAG retrieval.
+Business background knowledge provided as files, injected into the conversation context via RAG retrieval.
 
-Content suitable for the knowledge base:
+Suitable content for the knowledge base:
 
-* Business rule descriptions
-* Data dictionary / code mapping tables
+* Business rules
+* Data dictionaries / code mapping tables
 * Industry terminology explanations
-* Analysis methodology / calculation formula specifications
-* FAQ frequently asked questions
-* Content recalled only in specific scenarios
+* Analysis methodologies / calculation definitions
+* FAQ
+* Content that should only be recalled in specific scenarios
 
-**Core Value**: Helps the model understand the "business logic behind the data," not just the data structure.
+**Core value**: Helps the model understand the business logic behind the data, not just the data structure.
 
 ### 3. Metrics / Answer Builder
 
-Predefined calculation formulas supporting alias matching, ensuring consistent metric definitions. Specifically:
+Pre-defined calculation formulas with alias matching, ensuring consistent metric definitions. Specifically:
 
-* Metrics: Used to define aggregation calculations (e.g., SUM, AVG)
-* Answer Builder: Used to define multi-table JOINs and complex filter conditions
+* Metrics: used to define aggregate calculations (e.g., SUM, AVG)
+* Answer Builder: used to define queries involving multi-table JOINs and complex filter conditions
 
-| Element       | Description                                                                                           | Screenshot                                            |
-| ------------- | ----------------------------------------------------------------------------------------------------- | ----------------------------------------------------- |
-| Metric Name   | e.g., "Average Transaction Value"                                                                     | ![](/.topwrite/assets/image_1779160566375.png =215)   |
-| Calculation   | SUM(revenue) / COUNT(DISTINCT order\_id)                                                              | ![](/.topwrite/assets/image_1779160586259.png =206)   |
-| Alias         | Helps the model understand different names for the same concept                                       | ![](/.topwrite/assets/image_1779160603309.png =204)   |
-| Auto-Generation | Basic aggregate metrics (e.g., SUM, COUNT) can now be auto-generated for selected tables. Context-aware metric insights and auto-generation will be available in future releases, enabling smarter metric recommendations | ![](/.topwrite/assets/image_1779160824613.png =218)   |
+| Element               | Description                                                                                                                                                                                        | Screenshot demo                                     |
+| --------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------- |
+| Metric name           | e.g., "Average Order Value"                                                                                                                                                                        | ![](/.topwrite/assets/image_1780900229948.png =283) |
+| Calculation logic     | SUM(revenue) / COUNT(DISTINCT order\_id)                                                                                                                                                           | ![](/.topwrite/assets/image_1780900257971.png =155) |
+| Aliases               | Help the model understand different names for the same concept                                                                                                                                     | ![](/.topwrite/assets/image_1780900305727.png =178) |
+| Auto-generate metrics | Basic aggregate metrics (SUM, COUNT, etc.) can already be auto-generated for selected tables; context-aware metric insights and auto-generation for smarter metric recommendations are coming soon |                                                     |
 
-**Core Value**: Ensures key business metrics are calculated consistently and accurately, preventing different SQL from being generated each time
+**Core value**: Ensures key business metrics are calculated consistently and accurately — the same SQL is generated every time the same question is asked.
 
 ### 4. Domain Prompt
 
-System prompt at the analysis domain level, equivalent to giving the model a "role setting" and "behavioral guidelines."
+A system prompt at the analysis domain level, functioning as a "role definition" and "behavioral guidelines" for the model.
 
-* Role setting: "You are a data analyst for the Hong Kong food & beverage industry"
-* Answer guidelines: "Monetary amounts default to HKD, date format uses YYYY-MM-DD"
-* Business constraints: "When calculating same-store growth, only include stores that have been open for at least 12 months"
-* Output preferences: "For trend-related questions, prioritize using line charts"
+* Role definition: "You are a data analyst specializing in the Hong Kong food and beverage industry"
+* Response guidelines: "Currency defaults to HKD, date format is YYYY-MM-DD"
+* Business constraints: "When calculating same-store growth, include only stores that have been open for at least 12 months"
+* Output preferences: "For trend-related questions, prefer line charts"
 
-**Core Value**: Set global rules to avoid repeating prerequisites in every conversation
+**Core value**: Sets global rules so you do not have to repeat preconditions every conversation.
 
 ***
 
-## 2. Which Capability for Which Scenario? (Scenario → Capability Mapping)
+## 2. Which Capability Fits Which Scenario? (Scenario → Capability Mapping)
 
-| Scenario/Problem Type                       | Recommended Capability            | Why                                       |
-| -------------------------------------------- | --------------------------------- | ----------------------------------------- |
-| Model selects wrong table or column          | Table/Column Schema Annotations   | Unclear descriptions cause matching errors |
-| Filter values are wrong (e.g., using full name but data stores abbreviations) | Column Value Index                 | Establish fuzzy matching                  |
-| Calculation formula is wrong (e.g., "gross profit" algorithm is incorrect) | Metrics                          | Predefined exact formulas                  |
-| Complex calculation logic (multi-table JOIN + conditional filtering) | Answer Builder                   | Provides SQL templates                     |
-| Model does not understand business terms (e.g., "table turnover rate") | Knowledge Base                   | Provides terminology documentation         |
-| Background knowledge needed to answer (e.g., "why did March revenue decline") | Knowledge Base                   | Provides business context documentation    |
-| Need to repeat prerequisites every time      | Domain Prompt                     | Set once, effective globally               |
-| Same metric has multiple names               | Metric Alias / Column Alias       | Support matching with multiple expressions  |
+| Scenario / Problem Type                                                      | Recommended Capability           | Why                                            |
+| ---------------------------------------------------------------------------- | -------------------------------- | ---------------------------------------------- |
+| Many tables, model frequently picks wrong table                              | Partition analysis domains       | Narrows the model's selection space            |
+| Different teams need data isolation                                          | Partition analysis domains       | Domain-user binding provides natural isolation |
+| Model picks wrong table or column                                            | Table/column Schema descriptions | Unclear descriptions cause mismatches          |
+| Filter values are wrong (e.g., full name used but data stores abbreviations) | Column value index               | Enables fuzzy matching                         |
+| Calculation logic is wrong (e.g., "gross profit" computed incorrectly)       | Metrics                          | Pre-defines the exact formula                  |
+| Complex calculation logic (multi-table JOIN + conditional filters)           | Answer Builder                   | Provides a SQL template                        |
+| Model does not understand business terms (e.g., "table turnover rate")       | Knowledge base                   | Provides terminology explanation documents     |
+| Background knowledge needed to answer (e.g., "Why did March revenue drop?")  | Knowledge base                   | Provides business context documents            |
+| Preconditions have to be repeated every conversation                         | Domain prompt                    | Set once, applied globally                     |
+| Same metric has multiple names                                               | Metric alias / column alias      | Supports multiple expression matching          |
 
-### Decision Flowchart
+### Decision Flow
 
 ```
-
-User question is inaccurate
+User answer is inaccurate
 
 ↓
 
-Is it "wrong table/column selection"? → Yes → Improve Schema Annotations + Column Value Index
+Many tables, model frequently picks wrong table? → Yes → Partition analysis domains
 
 ↓ No
 
-Is it "wrong calculation logic"? → Yes → Create Metrics or Answer Builder
+Is it "wrong table/column"? → Yes → Complete Schema descriptions + column value index
 
 ↓ No
 
-Is it "not understanding business concepts"? → Yes → Add Knowledge Base documents
+Is it "wrong calculation logic"? → Yes → Create a Metric or Answer Builder
 
 ↓ No
 
-Is it "need to repeat prerequisites every time"? → Yes → Configure Domain Prompt
+Is it "does not understand business concepts"? → Yes → Add knowledge base documents
 
 ↓ No
 
-Investigate using a combination of the above methods
+Is it "preconditions must be repeated every time"? → Yes → Configure domain prompt
+
+↓ No
 
+Investigate using a combination of the above techniques
 ```
 
 ***
 
 ## 3. How to Use? (Knowledge Base Management)
 
-### 3.1 Knowledge Base Organization
+### 3.1 Knowledge Base Structure
 
-The new knowledge base uses a hierarchical folder structure (similar to Lark/Feishu knowledge base):
-![](/.topwrite/assets/image_1779161523552.png)
+The new knowledge base uses a hierarchical folder structure (similar to Feishu/Lark knowledge base):
 
 ### 3.2 Content Creation Methods
 
-| Method         | Applicable Scenarios                           |
-| -------------- | ----------------------------------------------- |
-| Create directly | Manually write business rules, terminology explanations, etc. |
-| Local upload   | Upload existing PDF/Word/Excel/Markdown documents |
-| Cloud import   | Batch import existing documents from OSS/S3     |
+| Method                    | Suitable Scenario                                             |
+| ------------------------- | ------------------------------------------------------------- |
+| Create directly           | Manually write business rules, terminology explanations, etc. |
+| Upload locally            | You already have PDF/Word/Excel/Markdown documents            |
+| Import from cloud storage | Bulk import existing documents from OSS/S3                    |
 
-### 3.3 Associate with Analysis Domains (Critical!)
+### 3.3 Associate with Analysis Domain (Critical!)
 
-Knowledge base content must be associated with an analysis domain for conversations under that domain to retrieve it:
+Knowledge base content must be associated with an analysis domain before conversations in that domain can retrieve it:
 
-* Can associate at the folder level → child files automatically inherit
-* A single file can be associated with multiple analysis domains
+* Associate at the folder level → sub-files inherit automatically
+* One file can be associated with multiple analysis domains
 * Unassociated files will not be retrieved by any conversation
 
-**Best Practices**:
+**Best practices**:
 
-* Create folders by business theme, associate at the folder level
-* General knowledge (e.g., company glossary) associates with all analysis domains
-* Specific business knowledge associates only with the corresponding domain
+* Create folders by business topic, associate domains at the folder level
+* Associate general knowledge (e.g., company glossary) with all analysis domains
+* Associate domain-specific knowledge only with the corresponding domain
 
-### 3.4 What Content is Suitable for Knowledge Base?
+### 3.4 What Content Belongs in the Knowledge Base?
 
-✅ Content suitable for knowledge base:
-
-| Content Type     | Example                                                    |
-| ---------------- | ---------------------------------------------------------- |
-| Business rules   | Table turnover rate = tables served / total tables; Average transaction value = revenue / customer count |
-| Code mapping     | Store code A01 = Causeway Bay store, A02 = Mong Kok store |
-| Industry terms   | Definitions of "dine-in", "takeaway", "table turnover", "preparation time" |
-| Data definitions | "Valid orders" exclude refunded orders and test orders     |
-| Analysis methods | Same-store growth only counts stores open for 12+ months   |
-| Common FAQ       | "Why don't the report numbers match finance?" → Because the report excludes tax |
-| Background info  | In Q1 2026, the company adjusted stores: 3 new openings, 2 closures |
-
-❌ Content NOT suitable for knowledge base:
+Content suitable for the knowledge base:
 
-| Content Type                | What to Use Instead           |
-| --------------------------- | ----------------------------- |
-| Exact calculation formulas  | → Use the Metrics feature     |
-| Complex SQL templates       | → Use the Answer Builder      |
-| Column aliases/descriptions | → Use Schema Annotations      |
-| Global answer guidelines    | → Use Domain Prompt           |
+| Content Type           | Example                                                                                            |
+| ---------------------- | -------------------------------------------------------------------------------------------------- |
+| Business rules         | Table turnover rate = tables served / total tables; average order value = revenue / customer count |
+| Code mapping           | Store code A01 = Causeway Bay store, A02 = Mong Kok store                                          |
+| Industry terminology   | Definitions of "dine-in", "delivery", "table turnover", "service time"                             |
+| Data definitions       | "Valid orders" exclude refunded orders and test orders                                             |
+| Analysis methodology   | Same-store growth counts only stores open for at least 12 months                                   |
+| Common FAQ             | "Why does report data not match accounting?" → Because the report excludes tax                     |
+| Background information | In Q1 2026 the company adjusted stores: 3 new openings, 2 closures                                 |
+
+Content not suitable for the knowledge base:
+
+| Content Type                           | What to Use Instead         |
+| -------------------------------------- | --------------------------- |
+| Exact calculation formulas for metrics | → Use the "Metrics" feature |
+| Complex SQL templates                  | → Use the "Answer Builder"  |
+| Column aliases and descriptions        | → Use "Schema Descriptions" |
+| Global response guidelines             | → Use the "Domain Prompt"   |
 
 ### 3.5 Management Recommendations
 
-| Recommendation          | Description                                           |
-| ----------------------- | ----------------------------------------------------- |
-| Organize by theme folder | Facilitates management and batch domain association  |
-| Keep documents concise  | Focus each file on one topic; avoid being overly broad |
-| Update regularly        | Sync knowledge base when business rules change        |
-| Naming conventions      | File name reflects content, e.g., "table-turnover-rate-rules.md" |
-| Verify before rollout   | Test a few questions after adding knowledge before broad use |
+| Recommendation              | Details                                                                                  |
+| --------------------------- | ---------------------------------------------------------------------------------------- |
+| Use folders by topic        | Easier to manage and bulk-associate with analysis domains                                |
+| Keep documents concise      | Each file focuses on one topic; avoid large all-in-one documents                         |
+| Update regularly            | Sync the knowledge base when business rules change                                       |
+| Use meaningful names        | File names should reflect content, e.g., "table-turnover-rate-calculation-rules.md"      |
+| Validate before rolling out | Test a few questions after adding knowledge to confirm the effect before wide deployment |
 
-## 4. Summary: Synergy of the 4 Capabilities
+## 4. Summary: How the 4 Capabilities Work Together
 
 In one sentence:
 
-* Schema Annotations help the model "find the right data"
+* Schema descriptions help the model "find the right data"
 * Metrics help the model "calculate the right numbers"
-* Knowledge Base helps the model "understand the business"
-* Domain Prompt helps the model "follow the rules"
+* Knowledge base helps the model "understand the business"
+* Domain prompt helps the model "follow the rules"
 
-Use all four together for the highest accuracy.
+Using all four together yields the highest accuracy.

package/bin/skills/lakehouse-doc-en/references/application-list.md

@@ -10,10 +10,8 @@ You must have already applied for resource permissions. For related operation in
 
 Click **Approval > Application** in the left menu to enter the **Application Records** page.
 
-![](.topwrite/assets/image_1736245885982.png)
-
 You can perform the following operations:
-![](.topwrite/assets/image_1736245926683.png)
+
 
 * **View work order details**: Click a **Work Order ID** in the list. In the pop-up side panel, you can view details of the work order, including application details, requested resources, approval status, etc.
 

package/bin/skills/lakehouse-doc-en/references/approval-list.md

@@ -13,21 +13,20 @@ Based on the approval flow settings for each resource, as an object permission o
 
 When a user applies for permissions, the system automatically creates an approval work order based on the target object's approval flow for review. The applicant can view their application information and approval status under **Approval > Application**.
 
-![](.topwrite/assets/image_1736245807432.png)
+^
 
 When an approver receives an approval notification, they can see pending work orders under **Approval > Approval**. Specific operations are as follows:
 
 * Click **Approval > Approval** in the left menu to enter the "Pending My Approval" tab.
 
 * The approval page supports both card view and list view.
-  ![](.topwrite/assets/image_1736245843874.png)
 
 * Filter conditions: Set keywords or work order ID, applicant, data object type, application time, sort order, and other search criteria to query pending approval work orders that meet the conditions. When multiple search conditions are set, the intersection of all conditions is used for the query.
 
 **1. View Work Order**
 
-*   **Card view**: Click a specific work order card on the left to view its application details on the right.
-*   **List view**: Click the ID column in the list to open a pop-up with the current approval work order details.
+* **Card view**: Click a specific work order card on the left to view its application details on the right.
+* **List view**: Click the ID column in the list to open a pop-up with the current approval work order details.
 
 Details include:
 
@@ -37,25 +36,25 @@ Details include:
 
 On the "Pending My Approval" page, approvers can perform the following approval operations:
 
-*   **Approve**
-    *   Click the **Approve** button in the **Operation** column of a work order in the list, or the **Approve** button on the work order details page, or the **Approve** button in the upper right corner of the card view. In the pop-up window, click **Confirm** to approve the application.
-    *   Select one or more items, then click the **Approve** button at the bottom of the list. In the pop-up window, click **Confirm** to approve applications in bulk.
+* **Approve**
+  * Click the **Approve** button in the **Operation** column of a work order in the list, or the **Approve** button on the work order details page, or the **Approve** button in the upper right corner of the card view. In the pop-up window, click **Confirm** to approve the application.
+  * Select one or more items, then click the **Approve** button at the bottom of the list. In the pop-up window, click **Confirm** to approve applications in bulk.
 
-*   **Reject**
-    *   Click the **Reject** button in the **Operation** column of a work order in the list, or the **Reject** button on the work order details page, or the **Reject** button in the upper right corner of the card view. In the pop-up window, enter the rejection reason and click **Confirm** to reject the application.
-    *   Select one or more items, then click the **Reject** button at the bottom of the list. In the pop-up window, click **Confirm** to reject applications in bulk.
+* **Reject**
+  * Click the **Reject** button in the **Operation** column of a work order in the list, or the **Reject** button on the work order details page, or the **Reject** button in the upper right corner of the card view. In the pop-up window, enter the rejection reason and click **Confirm** to reject the application.
+  * Select one or more items, then click the **Reject** button at the bottom of the list. In the pop-up window, click **Confirm** to reject applications in bulk.
 
-*   **Add Signer**: Indicates that after the current approver approves, additional signers are still needed to participate in approval.
-    *   Click the **Add Signer** button in the **Operation** column of a work order in the list, or the **Add Signer** button on the work order details page, or the **Add Signer** button in the upper right corner of the card view. In the **Add Signer** pop-up window, configure the signer information and click **Confirm** to complete the add-signer approval for the application.
-    *   Select one or more items, then click the **Add Signer** button at the bottom of the list. In the pop-up window, configure the signer information and click **Confirm** to complete batch add-signer.
+* **Add Signer**: Indicates that after the current approver approves, additional signers are still needed to participate in approval.
+  * Click the **Add Signer** button in the **Operation** column of a work order in the list, or the **Add Signer** button on the work order details page, or the **Add Signer** button in the upper right corner of the card view. In the **Add Signer** pop-up window, configure the signer information and click **Confirm** to complete the add-signer approval for the application.
+  * Select one or more items, then click the **Add Signer** button at the bottom of the list. In the pop-up window, configure the signer information and click **Confirm** to complete batch add-signer.
 
-**Note:**
+**Note**:
 
-*   When the approver clicks the add-signer operation, the system automatically treats the current approver as having performed the approval;
-*   When a work order contains multiple approval items, the approval operation applies to all items, meaning you can only approve all or reject all at once.
+* When the approver clicks the add-signer operation, the system automatically treats the current approver as having performed the approval;
+* When a work order contains multiple approval items, the approval operation applies to all items, meaning you can only approve all or reject all at once.
 
 ### Completed
 
 Click the "Completed" tab under the Approval function to view all approval work orders that you have processed and those awaiting your action within the current instance.
 
-![](https://tq2pllvokz.feishu.cn/space/api/box/stream/download/asynccode/?code=YmU2MjVkZTA4OWU3NDlkYWZmOGI1ZTZlOTFlZmY4MWVfZHAwaE94eWpkUzB2S2NWSGppcFd3NWsxekw3QmlFN3ZfVG9rZW46UWl5QWJ1SmZnb0pkZ1h4NDJUWmNGMUpJbjZmXzE3Mjk4MjU3MDQ6MTcyOTgyOTMwNF9WNA)
+^

package/bin/skills/lakehouse-doc-en/references/batch-load-parquet-file-into-lakehouse.md

@@ -12,7 +12,7 @@ This guide will help you import large amounts of data from public URL Parquet fi
 
 Script download address: <https://github.com/yunqiqiliang/nyc-taxi-data-clickzetta>
 
-## 1. Install [Singdata SQLLine](https://doc.clickzetta.com/zh-CN/connect-with-cli)
+## 1. Install [Singdata SQLLine](connect-with-cli.md)
 
 ## 2. Install [R](https://www.r-project.org/)