Merge pull request #11 from basicmachines-co/fix/remove-hardcoded-bm-local-path

fix: replace hardcoded bm_local_path with env var

Author: phernandez

.github/workflows/benchmark-nightly.yml

@@ -13,6 +13,8 @@ jobs:
 
       - uses: astral-sh/setup-uv@v4
 
+      - uses: extractions/setup-just@v2
+
       - name: Install dependencies
         run: uv sync --group dev --extra judge
 

.github/workflows/benchmark-smoke.yml

@@ -12,6 +12,8 @@ jobs:
 
       - uses: astral-sh/setup-uv@v4
 
+      - uses: extractions/setup-just@v2
+
       - name: Install dependencies
         run: uv sync --group dev
 

README.md

@@ -80,7 +80,7 @@ Local override:
 
 ```bash
 uv run bm-bench run retrieval \
-  --bm-local-path /Users/phernandez/dev/basicmachines/basic-memory
+  --bm-local-path /path/to/basic-memory
 ```
 
 ## Mem0 local requirements

docs/benchmarks.md

@@ -42,8 +42,8 @@ It covers:
 
 ### Repositories and paths
 
-- benchmark repo: `/Users/phernandez/dev/basicmachines/basic-memory-benchmarks`
-- BM local repo (default in `justfile`): `/Users/phernandez/dev/basicmachines/basic-memory`
+- benchmark repo: clone of `basicmachines-co/basic-memory-benchmarks`
+- BM local repo: set `BM_LOCAL_PATH` env var (or in `.env`) to your local `basic-memory` checkout
 
 ### Environment
 
@@ -53,7 +53,7 @@ It covers:
 ### One-time setup
 
 ```bash
-cd /Users/phernandez/dev/basicmachines/basic-memory-benchmarks
+cd /path/to/basic-memory-benchmarks
 just sync
 ```
 
@@ -107,7 +107,7 @@ Top-level commands:
 ### One-command full retrieval run
 
 ```bash
-cd /Users/phernandez/dev/basicmachines/basic-memory-benchmarks
+cd /path/to/basic-memory-benchmarks
 just bench-full
 ```
 
@@ -119,7 +119,7 @@ This runs:
 ### One-command full retrieval + judge
 
 ```bash
-cd /Users/phernandez/dev/basicmachines/basic-memory-benchmarks
+cd /path/to/basic-memory-benchmarks
 just bench-full-judge
 ```
 
@@ -235,8 +235,8 @@ Use this workflow today to compare BM revisions while keeping benchmark tooling
 ### Step 1: Create BM worktrees for target refs
 
 ```bash
-BM_REPO=/Users/phernandez/dev/basicmachines/basic-memory
-WT_ROOT=/Users/phernandez/dev/basicmachines/basic-memory-benchmarks/benchmarks/worktrees/basic-memory
+BM_REPO=/path/to/basic-memory
+WT_ROOT=/path/to/basic-memory-benchmarks/benchmarks/worktrees/basic-memory
 
 mkdir -p "$WT_ROOT"
 
@@ -250,7 +250,7 @@ git -C "$BM_REPO" worktree add "$WT_ROOT/current" HEAD
 ### Step 2: Prepare benchmark datasets once
 
 ```bash
-cd /Users/phernandez/dev/basicmachines/basic-memory-benchmarks
+cd /path/to/basic-memory-benchmarks
 just sync
 just bench-prepare-short
 just bench-prepare-long
@@ -377,7 +377,7 @@ Planned command shape:
 
 ```bash
 uv run bm-bench run revision-matrix \
-  --bm-repo-path /Users/phernandez/dev/basicmachines/basic-memory \
+  --bm-repo-path /path/to/basic-memory \
   --revisions pre_fusion=f5a0e942^ \
   --revisions fusion=f5a0e942 \
   --revisions context_step1=f9b2a075 \

justfile

@@ -4,7 +4,8 @@ set dotenv-load := true
 
 # --- Paths and defaults ---
 
-bm_local_path := "/Users/phernandez/dev/basicmachines/basic-memory"
+bm_local_path := env_var_or_default("BM_LOCAL_PATH", "")
+bm_local_path_flag := if bm_local_path != "" { "--bm-local-path " + bm_local_path } else { "" }
 locomo_dataset_path := "benchmarks/datasets/locomo/locomo10.json"
 locomo_output_dir := "benchmarks/generated/locomo"
 locomo_c1_output_dir := "benchmarks/generated/locomo-c1"
@@ -84,7 +85,7 @@ bench-run-short:
       --corpus-dir benchmarks/generated/locomo-c1/docs \
       --queries-path benchmarks/generated/locomo-c1/queries.quick25.json \
       --providers bm-local,mem0-local \
-      --bm-local-path {{bm_local_path}} \
+      {{bm_local_path_flag}} \
       --allow-provider-skip
 
 bench-run-short-strict:
@@ -94,7 +95,7 @@ bench-run-short-strict:
       --corpus-dir benchmarks/generated/locomo-c1/docs \
       --queries-path benchmarks/generated/locomo-c1/queries.quick25.json \
       --providers bm-local,mem0-local \
-      --bm-local-path {{bm_local_path}} \
+      {{bm_local_path_flag}} \
       --strict-providers
 
 # Long benchmark: full LoCoMo query set
@@ -105,7 +106,7 @@ bench-run-long:
       --corpus-dir benchmarks/generated/locomo/docs \
       --queries-path benchmarks/generated/locomo/queries.json \
       --providers bm-local,mem0-local \
-      --bm-local-path {{bm_local_path}} \
+      {{bm_local_path_flag}} \
       --allow-provider-skip
 
 bench-run-long-strict:
@@ -115,7 +116,7 @@ bench-run-long-strict:
       --corpus-dir benchmarks/generated/locomo/docs \
       --queries-path benchmarks/generated/locomo/queries.json \
       --providers bm-local,mem0-local \
-      --bm-local-path {{bm_local_path}} \
+      {{bm_local_path_flag}} \
       --strict-providers
 
 bench-run-bm-local:
@@ -125,7 +126,7 @@ bench-run-bm-local:
       --dataset-path {{locomo_dataset_path}} \
       --corpus-dir benchmarks/generated/locomo/docs \
       --queries-path benchmarks/generated/locomo/queries.json \
-      --bm-local-path {{bm_local_path}}
+      {{bm_local_path_flag}}
 
 bench-run-mem0-local:
     uv run bm-bench run retrieval \
@@ -143,7 +144,7 @@ bench-run-full:
       --corpus-dir benchmarks/generated/locomo/docs \
       --queries-path benchmarks/generated/locomo/queries.json \
       --providers bm-local,mem0-local \
-      --bm-local-path {{bm_local_path}} \
+      {{bm_local_path_flag}} \
       --allow-provider-skip
 
 bench-run-full-judge model="gpt-4o-mini":
@@ -153,7 +154,7 @@ bench-run-full-judge model="gpt-4o-mini":
       --corpus-dir benchmarks/generated/locomo/docs \
       --queries-path benchmarks/generated/locomo/queries.json \
       --providers bm-local,mem0-local \
-      --bm-local-path {{bm_local_path}} \
+      {{bm_local_path_flag}} \
       --allow-provider-skip \
       --judge \
       --judge-model "{{model}}"

pyproject.toml

@@ -8,7 +8,7 @@ authors = [
 ]
 requires-python = ">=3.12"
 dependencies = [
-    "basic-memory @ git+https://github.com/basicmachines-co/basic-memory@main",
+    "basic-memory>=0.20.2",
     "httpx>=0.28.0",
     "mcp>=1.23.1",
     "mem0ai==1.0.0",