🔥 chore: remove 3gpp-ai.toml configuration file

# 3GPP AI Configuration

# Generated by: 3gpp-ai config init

#

# Configuration precedence (highest to lowest):

# 1. CLI arguments

# 2. This config file

# 3. Environment variables (TDC_*, LIGHTRAG_*)

# 4. Hard-coded defaults

#

# Sensitive values use ${ENV_VAR} syntax for security.

# Environment variables are interpolated at runtime.

[path]

# Root cache directory for storing downloaded files and metadata

# Default: C:\Users\Jan.Reimes\.3gpp-crawler

cache_dir = "C:\Users\Jan.Reimes\.3gpp-crawler"

# SQLite database filename for storing crawl metadata

# Default: 3gpp_crawler.db

db_filename = "3gpp_crawler.db"

# Subdirectory name for checked-out documents

# Default: checkout

checkout_dirname = "checkout"

# Subdirectory name for AI-related cache (embeddings, graphs)

# Default: lightrag

ai_cache_dirname = "lightrag"

[http]

# Time-to-live for HTTP cache entries in seconds

# Default: 7200

cache_ttl = 7200

# Enable HTTP response caching

# Default: True

cache_enabled = true

# Refresh cache TTL on each access

# Default: True

cache_refresh_on_access = true

# Verify SSL certificates for HTTPS requests

# Default: True

verify_ssl = true

# Maximum number of retry attempts for failed requests

# Default: 3

max_retries = 3

# HTTP request timeout in seconds

# Default: 30

timeout = 60

[credentials]

# Username for ETSI Online (EOL) portal authentication

username = "reimes"

# Password for ETSI Online (EOL) portal authentication

password = "1y4RDXua9HOuUC"

# Custom prompt message for interactive credential entry

prompt = "false"

[crawl]

# Filter by working group (e.g., S4, RAN1, CT3)

working_group = null

# Filter by sub-working group

sub_group = null

# Start date filter (YYYY-MM-DD, YYYY-MM, or YYYY format)

date_start = null

# End date filter (YYYY-MM-DD, YYYY-MM, or YYYY format)

date_end = null

# SQL LIKE pattern to match document source

source_like = null

# SQL LIKE pattern to match agenda item

agenda_like = null

# SQL LIKE pattern to match document title

title_like = null

# Maximum number of documents to crawl

# Default: 1000

limit = 1000

# Number of concurrent workers for crawling

# Default: 4

workers = 4

[llm]

# LLM model name in <provider>/<model> format

# Default: openrouter/openrouter/free

model = "openai/glm-4.7"

# LLM API base URL

# Default: http://localhost:11434

api_base = "https://api.z.ai/api/coding/paas/v4"

# API key for LLM provider

# Environment: TDC_AI_LLM_API_KEY

api_key = "${TDC_AI_LLM_API_KEY}"

[embedding]

# Embedding model name in <provider>/<model> format

# Default: ollama/qwen3-embedding:0.6b

model = "ollama/embeddinggemma:latest"

# Embedding API base URL

# Default: http://localhost:11434

api_base = "http://localhost:11434"

# API key for embedding provider

api_key = null

[database]

# Storage backend to use (file or pg0)

# Default: file

backend = "file"

# pg0 instance name

# Default: 3gpp-crawler

pg0_instance_name = "3gpp-crawler"

# pg0 PostgreSQL port

# Default: 15432

pg0_port = 15432

# pg0 database name

# Default: tdoc

pg0_database = "tdoc"

[extraction]

# Enable extraction and indexing of table elements

# Default: True

tables = true

# Enable extraction and indexing of figure elements

# Default: True

figures = true

# Enable extraction and indexing of equation elements

# Default: True

equations = true

# Enable figure description generation with vision-capable models

# Default: True

figure_description = true

[workspace]

# Default workspace name

# Default: default

default_name = "default"

# Default query mode (naive, local, global, hybrid, mix, bypass)

# Default: hybrid

default_query_mode = "hybrid"

# Enable shared embedding storage across workspaces (deduplication)

# Default: False

shared_storage = false