MEDIUM: Hardcoded URLs and Configuration Values #8

New Issue

Koko210 · 2026-02-16T22:09:38+02:00

Koko210 commented

2026-02-16 22:09:38 +02:00

Many URLs and configuration values are hardcoded in the codebase, making it difficult to change deployment targets or API endpoints.

Where It Occurs

bot/globals.py - LLM endpoints (LLAMA_URL, LLAMA_AMD_URL)
bot/config.py - Web UI URLs
bot/utils/llm.py - Direct URL references
bot/api.py - API base URLs

Why This Is a Problem

Deployment Issues: Cannot deploy to different environments without code changes
Maintenance: Changing API endpoints requires finding all references
Testing: Cannot use mock endpoints for testing
Security: Secrets may be accidentally committed

What Can Go Wrong

Scenario 1: Switching to New LLM Provider

Team decides to switch from local LLM to OpenAI API
Developer searches for all LLM URL references
Misses one reference in utils/llm.py
Bot still tries to call local LLM in one code path
Bot fails randomly when that code path is executed

Scenario 2: Staging Environment

Developer sets up staging environment
Needs to use staging API endpoints
Has to modify 15+ files to change URLs
Accidentally commits staging URLs to production
Production bot now calls staging APIs

Proposed Fix

Move all configuration to config.yaml and load via ConfigManager.

Severity

MEDIUM - Hardcoded values make deployment difficult and error-prone.

Files Affected

bot/globals.py, bot/config.py, bot/utils/llm.py, bot/api.py, config.yaml

Many URLs and configuration values are hardcoded in the codebase, making it difficult to change deployment targets or API endpoints. ## Where It Occurs - bot/globals.py - LLM endpoints (LLAMA_URL, LLAMA_AMD_URL) - bot/config.py - Web UI URLs - bot/utils/llm.py - Direct URL references - bot/api.py - API base URLs ## Why This Is a Problem 1. Deployment Issues: Cannot deploy to different environments without code changes 2. Maintenance: Changing API endpoints requires finding all references 3. Testing: Cannot use mock endpoints for testing 4. Security: Secrets may be accidentally committed ## What Can Go Wrong ### Scenario 1: Switching to New LLM Provider 1. Team decides to switch from local LLM to OpenAI API 2. Developer searches for all LLM URL references 3. Misses one reference in utils/llm.py 4. Bot still tries to call local LLM in one code path 5. Bot fails randomly when that code path is executed ### Scenario 2: Staging Environment 1. Developer sets up staging environment 2. Needs to use staging API endpoints 3. Has to modify 15+ files to change URLs 4. Accidentally commits staging URLs to production 5. Production bot now calls staging APIs ## Proposed Fix Move all configuration to config.yaml and load via ConfigManager. ## Severity MEDIUM - Hardcoded values make deployment difficult and error-prone. ## Files Affected bot/globals.py, bot/config.py, bot/utils/llm.py, bot/api.py, config.yaml

Koko210 closed this issue

2026-02-16 22:16:33 +02:00

Koko210 reopened this issue

2026-02-16 22:17:02 +02:00

Koko210 commented

2026-02-16 23:34:31 +02:00

Closing as Invalid - All URLs are already environment-configurable, not hardcoded. In bot/globals.py lines 25-26 and 36, all URLs use os.getenv() with Docker service name defaults: LLAMA_URL, LLAMA_AMD_URL, and CHESHIRE_CAT_URL. This is the standard pattern for docker-compose deployments. The defaults are appropriate Docker service names that work in the compose network.

Koko210 closed this issue

2026-02-16 23:34:40 +02:00

Koko210 referenced this issue from a commit

2026-02-23 13:31:38 +02:00

fix(P1): 6 priority-1 bug fixes for autonomous engine and mood system

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: Koko210/miku-discord#8