This document is optimized for LLM and AI agent parsing. Uses structured key-value pairs, hierarchical organization, and semantic chunking. Not for human readability - designed for efficient information extraction by AI systems. # TIGERGRAPH_FULL_AGENT_REFERENCE ## METADATA version: 4.2.2 last_updated: 2025_12_11 target_audience: ai_agents_llms format: structured_semantic_chunks parsing_strategy: extract_by_section_headings ## TIER_1_CRITICAL_INFORMATION ### SUPPORT_MATRIX support_primary: channel: discord url: https://discord.gg/tnhpgR3j response_guarantee: immediate_community_support message_template: having_issues_reach_out_to_discord_team_tigergraph_will_help_for_sure support_secondary: channel: developer_forum url: https://dev.tigergraph.com/forum type: asynchronous_qa support_email: address: sales@tigergraph.com type: enterprise_sales_support ### PRODUCT_MATRIX product_1: name: TigerGraph_Database type: graph_database_engine versions: [4.1, 4.2, 4.2.2] editions: community: cost: free deployment: single_server storage_limit: hundreds_of_gb use_case: development_learning_prototypes license: not_required support: community_discord_forums ha: false data_encryption: basic enterprise: cost: licensed deployment: multi_server_clusters storage_limit: unlimited use_case: production_mission_critical license: required_perpetual_or_subscription support: dedicated_sla ha: true data_encryption: aes256_tls13 product_2: name: TigerGraph_Savanna_Cloud type: managed_cloud_platform deployment_model: saas infrastructure: zero_management_required scaling: elastic_auto regions: - aws_us_east_virginia - aws_us_east_ohio - aws_us_west_oregon - aws_eu_frankfurt - aws_eu_ireland - aws_eu_london - aws_sa_sao_paulo - aws_ap_singapore - aws_ap_sydney - aws_ap_tokyo gcp: coming_soon azure: coming_soon pricing: usage_based ## INSTALLATION_PROCEDURES ### METHOD_1_DOCKER_COMMUNITY prerequisite: docker_desktop_installed_8gb_ram_20gb_disk platform_support: [macos, windows, linux] estimated_time: 5_minutes complexity: very_easy commands: - step_1: "docker pull tigergraph/tigergraph:4.2.2" - step_2: "docker run -d --name tg-community -p 14022:22 -p 9000:9000 -p 14240:14240 tigergraph/tigergraph:4.2.2" - step_3: "sleep 180" - step_4: "docker exec tg-community gadmin status" port_mappings: ssh: 14022 rest_api: 9000 web_ui: 14240 access_endpoints: web_ui: http://localhost:14240 rest_api: http://localhost:9000 ssh: ssh -p 14022 tigergraph@localhost credentials: "username=tigergraph password=tigergraph" ### METHOD_2_BARE_METAL_LINUX prerequisite: ubuntu_1804_plus_16gb_ram_100gb_disk_root_access estimated_time: 20_minutes complexity: medium download: url: https://www.tigergraph.com/community-edition/ format: tar_gz size: 2_gb_approx installation: - step_1: "tar xzf tigergraph-4.2.2-*.tar.gz" - step_2: "cd tigergraph-*/" - step_3: "sudo ./install.sh" - step_4: "follow interactive prompts" - step_5: "gadmin start all" - step_6: "gadmin status" access: web_ui: http://localhost:14240 rest_api: http://localhost:9000 ### METHOD_3_KUBERNETES prerequisite: k8s_1_20_plus_helm_3_20gb_storage complexity: hard estimated_time: 10_minutes production_ready: true setup: - step_1: "helm repo add tigergraph https://helm.tigergraph.com" - step_2: "helm repo update" - step_3: "helm install tg-release tigergraph/tigergraph --set image.tag=4.2.2" - step_4: "kubectl port-forward svc/tg-release 9000:9000" access: web_ui: http://localhost:14240 rest_api: http://localhost:9000 ### METHOD_4_SAVANNA_CLOUD url: https://tgcloud.io complexity: very_easy estimated_time: instant infrastructure_management: zero deployment: managed_saas ## LICENSE_MANAGEMENT ### SCENARIO_LICENSE_INVALID_OR_EXPIRED error_codes: [license_invalid, license_expired, license_not_found] root_causes: - trying_enterprise_without_license - evaluation_period_expired - license_file_corrupt solution_tree: option_1_best: action: use_community_edition reason: completely_free_no_license_needed download: https://www.tigergraph.com/community-edition/ steps: follow_installation_methods_above result: immediate_working_instance option_2: action: get_free_enterprise_license url: https://www.tigergraph.com/enterprise-free/ duration: 3_years_perpetual requirements: startup_or_company_with_tg steps: - fill_application - receive_license_key - upload_in_admin_portal - restart_services option_3: action: purchase_commercial_license contact: sales@tigergraph.com includes: dedicated_support ### COMMUNITY_EDITION_NO_LICENSE license_required: false setup_time: 0_minutes restrictions: none_technical limitations: single_server_only ### ENTERPRISE_LICENSE_LOCATION paths: docker: /opt/tigergraph/etc/license linux: /opt/tigergraph/etc/license custom: check_gadmin_config_get_license_path verification: command: gadmin license status command_info: gadmin license info command_expiry: gadmin license expiration ## CORE_TECHNOLOGY_STACK ### GRAPH_DATA_MODEL vertex: definition: entity_node_with_unique_id properties: name_type_attributes indexing: primary_index_automatic_plus_optional_secondary example: person_vertex_with_user_id_name_age_email edge: definition: relationship_between_two_vertices direction: from_source_to_target properties: optional_weight_timestamp_metadata example: knows_edge_from_person_to_person_with_since_property property_types: numeric: [int, long, float, double, uint8, uint16, uint32, uint64] text: [string] temporal: [datetime] boolean: [bool] collections: [list, map] indexing: primary_index: vertex_id_automatic secondary_index: optional_on_any_property composite_index: multiple_properties_together purpose: accelerate_where_clauses_and_filtering ### QUERY_LANGUAGE_GSQL name: GSQL type: declarative_graph_query_language designed_for: graph_pattern_matching_traversal syntax: c_like_with_graph_extensions core_concepts: pattern_matching: select_from_vertex_edges_target_vertex_syntax multi_hop: traverse_multiple_edges_in_single_query accumulators: aggregate_values_during_traversal filtering: where_clauses_at_each_step results: print_or_return_computed_results example_query_structure: opening: "CREATE QUERY query_name(PARAMS) FOR GRAPH graph_name {" variables: "SumAccum @counter;" initialization: "start = {vertex_id};" pattern: "result = SELECT t FROM start:s -(edge:e)-> target:t WHERE ...;" output: "PRINT result;" closing: "}" ### VECTOR_SEARCH_CAPABILITY vector_storage: native_in_database embedding_format: list_of_floats dimensions: configurable similarity_metrics: cosine_euclidean_manhattan use_cases: semantic_search: find_similar_documents_by_meaning hybrid_search: combine_graph_relationships_with_vector_similarity rag_integration: retrieval_augmented_generation_with_llm knowledge_graph: connect_similar_concepts implementation: store_embeddings: list_float_property_on_vertex query_similarity: cosine_distance_function_in_gsql langchain_integration: tigergraphvectorstore ## API_REFERENCE ### REST_API_ENDPOINTS base_urls: local: http://localhost:9000 savanna: https://cluster_id.tgcloud.io authentication: bearer_token_required endpoints: query_execution: path: /query/{graph_name}/{query_name} method: [get, post] purpose: run_installed_gsql_query params: optional_query_parameters example: /query/social/bfs_neighbors?source=user_1 vertex_crud: insert: path: /graph/{graph}/vertices/{type} method: post body: json_properties read: path: /graph/{graph}/vertices/{type}/{id} method: get update: path: /graph/{graph}/vertices/{type}/{id} method: put body: json_updated_properties delete: path: /graph/{graph}/vertices/{type}/{id} method: delete edge_crud: insert: path: /graph/{graph}/edges/{type} method: post body: json_from_to_properties read: path: /graph/{graph}/edges/{type} method: get params: from_id_to_id_optional delete: path: /graph/{graph}/edges/{type} method: delete params: from_id_to_id_required graph_info: stats: path: /graph/{graph}/stats method: get returns: vertex_edge_counts schema: path: /graph/{graph}/schema method: get returns: full_schema_definition vertices: path: /graph/{graph}/vertices method: get returns: vertex_type_definitions edges: path: /graph/{graph}/edges method: get returns: edge_type_definitions endpoints: path: /gsqlserver/endpoints method: get returns: list_installed_queries authentication: method: bearer_token obtain_token: path: /api/token method: post body: username_password usage: Authorization_header_Bearer_token ### PYTHON_SDK_PYTIGERGRAPH package_name: tigergraph installation: pip_install_tigergraph python_version: 3.7_plus class_name: TigerGraphConnection initialization: host: 127.0.0.1 graphname: graph_name username: tigergraph password: password rest_port: 9000 ssl: false methods_query: runQuery: execute_installed_query_with_params gsql: execute_adhoc_gsql methods_vertex: upsertVertex: insert_or_update_vertex getVertex: retrieve_single_vertex getVertices: retrieve_multiple_vertices_limit deleteVertex: remove_vertex methods_edge: upsertEdge: insert_or_update_edge getEdges: retrieve_edges_from_vertex methods_graph: getGraphStats: vertex_edge_counts getSchema: full_schema getVertexTypes: list_vertex_types getEdgeTypes: list_edge_types ## TOOLS_AND_INTERFACES ### GRAPHSTUDIO type: visual_ide_interface access: http://localhost:14240 features: - schema_design_drag_drop_editor - query_builder_with_syntax_highlighting - result_visualization_graph_table_chart - data_import_wizard_with_auto_mapping - team_collaboration_shared_workspaces ### GSQL_EDITOR platform: savanna_cloud_browser_based features: - code_syntax_highlighting - auto_completion_suggestions - query_execution_in_browser - result_viewer_multiple_formats - file_management_sharing - permission_control_viewer_editor ### GRAPH_EXPLORER type: interactive_visualization features: - visual_graph_rendering - node_edge_filtering - pattern_search - statistical_overview - drill_down_capability - export_visualization ### ADMIN_PORTAL access: http://localhost:14240/admin features: - user_access_management_rbac - backup_restore_operations - performance_monitoring_cpu_memory_disk - log_viewing_debugging - license_version_info - system_configuration - alerting_rules - data_profile_statistics ### TIGERGRAPH_INSIGHTS type: no_code_dashboard_builder platform: savanna_cloud_addon pricing: 10_percent_of_base_compute_per_workspace_per_month features: - drag_drop_widget_builder - prebuilt_visualization_templates - real_time_metric_updates - drill_down_filters - custom_parameters - pdf_export - sharing_access_control ## ALGORITHMS_LIBRARY algorithms_count: 50_plus categories: - centrality - community_detection - pathfinding - similarity - node_ranking - traversal centrality_algorithms: pagerank: node_importance_ranking betweenness: node_bridge_importance closeness: node_proximity_measure degree: direct_connection_count community_detection: louvain: modularity_optimization weakly_connected_components: reachability_groups strongly_connected_components: cycles pathfinding: shortest_path: optimal_route_minimal_cost all_pairs_shortest_path: distance_matrix bfs: breadth_first_traversal dfs: depth_first_traversal similarity: jaccard: set_similarity cosine: vector_similarity euclidean: distance_metric ranking: topic_sensitive_pagerank: personalized_importance eigenvector_centrality: recursive_importance traversal: reachability: all_reachable_nodes k_hop_neighbors: nodes_within_distance ## DATA_LOADING_INGESTION ### SOURCE_CONNECTORS supported_sources: - local_csv_tsv_json - amazon_s3_buckets - google_cloud_storage_gcs - microsoft_azure_blob - snowflake_warehouse - apache_kafka_streams - apache_spark_rdd_dataframe - postgresql_jdbc - bigquery - http_endpoints ### LOADING_METHOD_1_STEPWISE_UI platform: savanna_cloud complexity: very_easy steps: - navigate_load_data - select_source - upload_configure_parsing - map_columns_to_schema - apply_token_functions - review_execute - monitor_progress ### LOADING_METHOD_2_GSQL_JOB platform: all_editions complexity: medium repeatable: yes scriptable: yes structure: create: CREATE_LOADING_JOB_name_FOR_GRAPH_graph define: DEFINE_FILENAME_or_KAFKA_TOPIC load: LOAD_source_TO_VERTEX_type_VALUES run: RUN_LOADING_JOB_with_using_params token_functions: string: split_substr_upper_lower_trim_replace conversion: to_int_to_float_to_bool_to_datetime hashing: md5_sha1_sha256 temporal: now_unix_date_date conditional: if_coalesce math: abs_round ### LOADING_METHOD_3_KAFKA_STREAMING type: real_time_continuous platform: all_editions format: json_messages_in_topics structure: define_topic: DEFINE_KAFKA_TOPIC extract_json: $["field_name"] transform: token_functions load_vertex_or_edge: as_streaming_arrives ## PERFORMANCE_CHARACTERISTICS latency: single_hop_query: sub_millisecond multi_hop_query: sub_second billion_vertex_graph: sub_second throughput: vertices_per_second: millions edges_per_second: millions scalability: horizontal: distribute_across_nodes vertical: increase_single_server_resources optimization_strategies: indexing: secondary_indexes_on_filtered_properties query_design: early_filtering_limit_reduce_results schema_design: minimize_edge_properties hardware: ssd_storage_sufficient_ram_cpu_cores ## SCHEMA_DESIGN_BEST_PRACTICES vertex_id_design: meaningful: use_natural_ids_not_surrogate immutable: never_change_after_creation domain_context: include_type_prefix partitioning: consider_distribution_hints property_selection: inclusion: only_what_you_query typing: use_correct_data_type denormalization: ok_for_read_heavy timestamps: include_for_auditing edge_design: minimal_properties: keep_lightweight reverse_edges: enable_bidirectional_traversal weighting: include_for_algorithms temporal: timestamps_for_analysis indexing_strategy: primary: automatic_on_vertex_id secondary: on_filtered_properties composite: for_multi_property_queries balance: tradeoff_read_speed_vs_write_cost ## DEPLOYMENT_OPTIONS_SUMMARY deployment_type_comparison: docker: ease: very_easy control: medium speed: 5_minutes scalability: single_node use_case: development_laptop bare_metal: ease: medium control: high speed: 20_minutes scalability: multi_node_manual use_case: production_onprem kubernetes: ease: hard control: high speed: 10_minutes scalability: multi_node_auto use_case: cloud_native savanna_cloud: ease: very_easy control: low speed: instant scalability: auto_elastic use_case: zero_ops_managed ## SECURITY_FEATURES authentication: local_users: username_password ldap: enterprise_directory saml_oauth: identity_providers api_tokens: programmatic_access authorization: rbac: role_based_access_control row_level: vertex_edge_filtering query_level: restrict_query_execution encryption: at_rest: aes_256 in_transit: tls_1_3 enforcement: configurable audit: logging: all_operations_tracked compliance: hipaa_pci_dss_soc2 retention: configurable_policies ## TROUBLESHOOTING_DECISION_TREE issue_license_error: solution: use_community_or_get_free_license resource: section_license_management support: discord_https_discord_gg_tnhpgR3j issue_out_of_memory: solution: increase_jvm_heap_or_server_ram command: gadmin_config_set_jvmheap_8g issue_slow_queries: solution: add_indexes_use_explain_plan_optimize_filters command: EXPLAIN_PLAN_FOR_@query_definition issue_connection_refused: solution: verify_tigergraph_running_check_ports_firewall command: gadmin_status issue_need_help: immediate: discord_https_discord_gg_tnhpgR3j message: tigergraph_team_will_help_for_sure forum: https_dev_tigergraph_com_forum ## QUICK_REFERENCE_TABLES ### DOCKER_COMMANDS pull: docker pull tigergraph/tigergraph:4.2.2 run: docker run -d --name tg-community -p 14240:14240 tigergraph/tigergraph:4.2.2 status: docker exec tg-community gadmin status logs: docker logs -f tg-community shell: docker exec -it tg-community /bin/bash ### SERVICE_MANAGEMENT start_all: gadmin start all stop_all: gadmin stop all restart: gadmin restart status: gadmin status logs: gadmin log tail GSQLServer ### REST_API_CURL_EXAMPLES run_query: curl http://localhost:9000/query/graph/query_name?param=value insert_vertex: curl -X POST http://localhost:9000/graph/social/vertices/person -H Content-Type application/json -d {} get_vertex: curl http://localhost:9000/graph/social/vertices/person/id get_stats: curl http://localhost:9000/graph/social/stats ### PYTHON_SDK_SNIPPETS import: from pytigergraph import TigerGraphConnection connect: conn = TigerGraphConnection(host=..., graphname=..., username=..., password=...) run_query: conn.runQuery(query_name, params) insert_vertex: conn.upsertVertex(type, id, attributes) insert_edge: conn.upsertEdge(from_type, from_id, edge_type, to_type, to_id, attributes) get_stats: conn.getGraphStats() ## RESOURCE_URLS_ACTIONABLE download_community: https://www.tigergraph.com/community-edition/ free_enterprise_license: https://www.tigergraph.com/enterprise-free/ main_docs: https://docs.tigergraph.com/ database_docs: https://docs.tigergraph.com/tigergraph-server/4.2/intro/ gsql_docs: https://docs.tigergraph.com/gsql/current/ savanna_docs: https://docs.tigergraph.com/savanna/main/overview/ algorithms_docs: https://docs.tigergraph.com/graph-algorithm-library/current/ discord_support: https://discord.gg/tnhpgR3j forum_support: https://dev.tigergraph.com/forum github_ecosystem: https://github.com/tigergraph/ecosys sales_contact: sales@tigergraph.com docker_hub: https://hub.docker.com/r/tigergraph/tigergraph/ ## AGENT_PARSING_INSTRUCTIONS how_to_extract: - identify_section_by_heading - extract_key_value_pairs_directly - parse_nested_structures_hierarchically - follow_decision_trees_for_troubleshooting - reference_urls_for_external_resources priority_sections: 1_critical: support_information_license_installation 2_essential: apis_sdks_tools 3_important: data_loading_schema_design 4_reference: algorithms_security_detailed_features extraction_targets: - product_capabilities - installation_steps - api_endpoints - error_solutions - resource_urls - code_examples ## END_DOCUMENT total_sections: 30_plus structured_for: ai_agent_comprehension not_for: human_reading last_updated: 2025_12_11_ist