Feat/docker local postgres#53
Open
rglauco wants to merge 8 commits intowillchen96:mainfrom
Open
Conversation
Adds an optional AUTH_MODE=local that replaces Supabase with: - Local PostgreSQL (via pg) for all data storage - Custom JWT auth (bcrypt + jsonwebtoken) via /auth/login and /auth/register - PgAdapter: a thin Supabase-compatible query builder over pg.Pool - DbClient abstraction so all route files work in both modes unchanged Infrastructure: - backend/Dockerfile and frontend/Dockerfile for containerised builds - compose.yml: frontend + backend + postgres with named volumes for persistence - docker/postgres/init.sql: full schema without Supabase-specific auth tables - .env.example at repo root with compose variables Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@opennextjs/cloudflare peer dep constraint doesn't exactly match the pinned next@16.0.3, causing npm ci to fail without the flag. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
AuthContext reads localStorage once on mount. After localSignIn() stores the JWT, the context still had user=null, so the layout redirected back to /login. Fix: expose refreshAuth() in the context and call it immediately after localSignIn/localSignUp so the state is updated before router.push(). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
createClient("", "") throws when NEXT_PUBLIC_SUPABASE_URL is not set,
crashing any page that imports supabase. Two fixes:
1. supabase.ts: use dummy localhost URL when Supabase is not configured
so the module loads safely; add getSessionToken() that returns the
local JWT or Supabase session token based on NEXT_PUBLIC_AUTH_MODE.
2. Replace all supabase.auth.getSession() calls in hooks and components
with getSessionToken() so they work in both auth modes.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Three bugs in local/Docker mode:
1. PgAdapter.maybeSingle() returned the full rows array instead of a
single row because _run() only checked _singleMode, not
_maybeSingleMode. All routes using maybeSingle() got an array back
and treated it as "not found" → "Chat not found" on every POST /chat.
2. JavaScript arrays passed to pg as query parameters are formatted as
PostgreSQL array literals {a,b} which is invalid for jsonb columns.
Fixed in PgAdapter insert/update to JSON.stringify arrays before
binding.
3. Plain strings inserted into jsonb columns (chat_messages.content for
user messages) are not valid JSON, causing the INSERT to fail
silently. Fixed by JSON.stringify-ing string content in chat.ts,
projectChat.ts and tabular.ts before the INSERT.
Also added missing `workflow jsonb` column to chat_messages in both
init.sql and schema.sql — the backend was already inserting this field.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…r model for Ollama
Two bugs:
1. PgAdapter upsert with only the conflict column (e.g. upsert({ user_id })
on conflict user_id) produced "DO UPDATE SET " with an empty SET
clause — invalid SQL. Now falls through to DO NOTHING when updateCols
is empty, and supports the Supabase ignoreDuplicates:true option.
This fixed "syntax error at end of input" on ensureProfileRow calls.
2. resolveTitleModel checked "apiKeys.ollama !== undefined" to detect
Ollama, but the apiKeys object always initialises ollama: null (not
undefined), so the condition was always true — every user with no cloud
keys got title_model = local-gemma-26b which doesn't exist on Ollama.
Fixed to check hasEnvOllama() || key.trim(), and when Ollama is the
provider, reuse the user's configured tabular model (which they know is
installed) instead of a hardcoded default.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
@rglauco You sure you want to be uploading those pdfs? (Am just passing by this repo and haven't looked deep so many this is ok...) |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.