From 49adbe26ec113ab0d2b3f59d86d5f91b34420bf4 Mon Sep 17 00:00:00 2001
From: Petr <petr@keboola.com>
Date: Sun, 15 Mar 2026 00:59:45 +0100
Subject: [PATCH] Move server venv setup to bootstrap, remove cross-platform
 pip freeze sync

Server venv is created during bootstrap via SSH (same package list,
installed natively on Linux). Removes sync_data.sh section that copied
pip freeze output across platforms (Windows/macOS freeze is incompatible
with Linux).
---
 docs/setup/bootstrap.yaml |  7 ++++++-
 scripts/sync_data.sh      | 20 +++-----------------
 2 files changed, 9 insertions(+), 18 deletions(-)

diff --git a/docs/setup/bootstrap.yaml b/docs/setup/bootstrap.yaml
index d98d3ab..19d942b 100644
--- a/docs/setup/bootstrap.yaml
+++ b/docs/setup/bootstrap.yaml
@@ -51,11 +51,16 @@ steps:
       - "rsync -avz --no-perms --no-group --progress {ssh_alias}:server/parquet/ ./server/parquet/"
     note: "Some folders may be empty if data sync hasn't run on the server yet. That's OK."
 
-  - name: "Set up Python venv"
+  - name: "Set up Python venv (local and server)"
+    description: |
+      Set up local venv, then create a matching venv on the server via SSH.
+      The server venv is needed for running scripts remotely (notifications, etc.).
+      Install the SAME packages on both sides (do NOT copy pip freeze across platforms).
     commands:
       - "python3 -m venv .venv"
       - "source .venv/bin/activate"
       - "pip install pandas pyarrow duckdb pyyaml python-dotenv"
+      - "ssh {ssh_alias} 'python3 -m venv ~/.venv && ~/.venv/bin/pip install --quiet pandas pyarrow duckdb pyyaml python-dotenv'"
 
   - name: "Initialize DuckDB"
     condition: "only if server/scripts/setup_views.sh exists"
diff --git a/scripts/sync_data.sh b/scripts/sync_data.sh
index 6cd4bc5..6023c07 100755
--- a/scripts/sync_data.sh
+++ b/scripts/sync_data.sh
@@ -491,23 +491,9 @@ else
     echo "  [dry-run] Would sync corporate memory rules to .claude/rules/"
 fi
 
-# Sync Python environment to server (only if dependencies changed)
-if [[ -z "$DRY_RUN" ]] && [[ -f "./.venv/bin/pip" ]]; then
-    echo ""
-    LOCAL_REQ="$(mktemp)"
-    ./.venv/bin/pip freeze > "$LOCAL_REQ"
-    LOCAL_HASH=$(md5sum "$LOCAL_REQ" 2>/dev/null | cut -d' ' -f1 || md5 -q "$LOCAL_REQ" 2>/dev/null || echo "none")
-    REMOTE_HASH=$(ssh $SSH_HOST "md5sum ~/.analyst_requirements.txt 2>/dev/null | cut -d' ' -f1 || echo 'missing'" 2>/dev/null || echo "missing")
-    if [[ "$LOCAL_HASH" != "$REMOTE_HASH" ]]; then
-        echo "Syncing Python environment to server..."
-        scp "$LOCAL_REQ" "$SSH_HOST:~/.analyst_requirements.txt"
-        ssh $SSH_HOST "test -d ~/.venv || python3 -m venv ~/.venv; ~/.venv/bin/pip install -r ~/.analyst_requirements.txt --quiet 2>&1 | tail -1"
-        echo "Server Python environment synced"
-    else
-        echo "Server Python environment up to date"
-    fi
-    rm -f "$LOCAL_REQ"
-fi
+# Note: Python environment on server is set up during bootstrap (not synced).
+# Server venv is created via: ssh {alias} 'python3 -m venv ~/.venv && pip install ...'
+# This avoids cross-platform pip freeze issues (Windows/macOS -> Linux).
 
 # Only update DuckDB and check freshness if NOT dry-run
 if [[ -z "$DRY_RUN" ]]; then