SourceIndex 0.1.0__tar.gz → 0.1.2__tar.gz

{sourceindex-0.1.0 → sourceindex-0.1.2}/.gitignore

@@ -13,3 +13,7 @@ venv/
 node_modules/
 .wrangler/
 .pytest_cache/
+
+# LLM-judge run outputs under tests/multilang/ — regenerable artifacts,
+# never committed (same policy as inspect/logs/).
+tests/multilang/results/

sourceindex-0.1.2/GETTING_STARTED.md

@@ -0,0 +1,83 @@
+# Getting Started with SourceIndex
+
+Hey — and thank you! You're the very first person outside the team to try
+SourceIndex, and that genuinely means a lot. I'm really glad you're giving
+it a go.
+
+Anything you notice along the way — good, bad, broken, or just "huh, that's
+weird" — would be hugely helpful to hear. No feedback is too small.
+
+## What it is
+
+When you ask a coding agent to fix or build something, it has to find the
+relevant code first. By default Claude Code and opencode do that live, in
+your session. SourceIndex does it ahead of time so the agent jumps straight
+to the right files.
+
+**Best on an existing repo you didn't fully write yourself.** Not very useful
+on a brand-new empty project.
+
+## Install and set up
+
+You'll need Python 3.10+ and a SourceIndex API key (starts with `sk-si-`).
+Apply for access at https://sourceindex.dev/access.
+
+**Easiest path:** open this file in Claude Code or opencode and tell it "set
+sourceindex up in this repo." It can run the steps for you.
+
+If you'd rather do it by hand:
+
+```bash
+pip install sourceindex
+cd /path/to/your/repo
+sourceindex init             # auto-detects your agent(s)
+sourceindex init claude      # set up Claude Code only
+sourceindex init opencode    # set up opencode only
+```
+
+It'll prompt for the key, set up some git hooks plus a subagent for the
+coding agent your repo already uses (a `CLAUDE.md` or `.claude/` means
+Claude Code; an `AGENTS.md` or `.opencode/` means opencode; neither means
+Claude Code), and scan your repo. A few minutes, once per repo. You can
+also name both at once (`sourceindex init claude opencode`), and it's safe
+to re-run init later to add the other agent.
+If you don't have a `CLAUDE.md` (Claude Code) or `AGENTS.md` (opencode)
+yet, run `/init` in your agent first.
+
+To replace a rotated or expired key later:
+
+```bash
+sourceindex api-key set
+```
+
+That prompts for the new key and persists it as your user-level default. New
+repos and repos without a local override will use it automatically. For
+automation, use
+`sourceindex api-key set --from-env` with `SOURCEINDEX_API_KEY` already set,
+pass `--api-key sk-si-...`, or pass the key positionally:
+`sourceindex api-key set sk-si-...`.
+
+To set a repo-local override:
+
+```bash
+sourceindex api-key set --local
+```
+
+Explicit environment variables still win, and a repo-local key overrides the
+global default.
+
+## Day-to-day
+
+Nothing. Open the repo in Claude Code or opencode as usual — it'll use
+SourceIndex on its own.
+
+## Feedback I'd love
+
+- Did `sourceindex init` finish cleanly?
+- Does the agent feel different? Even "I can't tell" is useful.
+- Anything in this doc that didn't land.
+
+More about the project lives at https://sourceindex.dev, and you can reach
+us any time at hi@sourceindex.dev.
+
+Thanks!

sourceindex-0.1.2/PKG-INFO

@@ -0,0 +1,103 @@
+Metadata-Version: 2.4
+Name: SourceIndex
+Version: 0.1.2
+Summary: Codebase index for agentic coding
+Requires-Python: >=3.10
+Requires-Dist: cryptography>=42
+Requires-Dist: keyring>=24
+Requires-Dist: litellm>=1.50
+Requires-Dist: tree-sitter-c>=0.23
+Requires-Dist: tree-sitter-cpp>=0.23
+Requires-Dist: tree-sitter-go>=0.23
+Requires-Dist: tree-sitter-java>=0.23
+Requires-Dist: tree-sitter-javascript>=0.23
+Requires-Dist: tree-sitter-kotlin>=1.0
+Requires-Dist: tree-sitter-ruby>=0.23
+Requires-Dist: tree-sitter-rust>=0.23
+Requires-Dist: tree-sitter-typescript>=0.23
+Requires-Dist: tree-sitter>=0.23
+Description-Content-Type: text/markdown
+
+# Getting Started with SourceIndex
+
+Hey — and thank you! You're the very first person outside the team to try
+SourceIndex, and that genuinely means a lot. I'm really glad you're giving
+it a go.
+
+Anything you notice along the way — good, bad, broken, or just "huh, that's
+weird" — would be hugely helpful to hear. No feedback is too small.
+
+## What it is
+
+When you ask a coding agent to fix or build something, it has to find the
+relevant code first. By default Claude Code and opencode do that live, in
+your session. SourceIndex does it ahead of time so the agent jumps straight
+to the right files.
+
+**Best on an existing repo you didn't fully write yourself.** Not very useful
+on a brand-new empty project.
+
+## Install and set up
+
+You'll need Python 3.10+ and a SourceIndex API key (starts with `sk-si-`).
+Apply for access at https://sourceindex.dev/access.
+
+**Easiest path:** open this file in Claude Code or opencode and tell it "set
+sourceindex up in this repo." It can run the steps for you.
+
+If you'd rather do it by hand:
+
+```bash
+pip install sourceindex
+cd /path/to/your/repo
+sourceindex init             # auto-detects your agent(s)
+sourceindex init claude      # set up Claude Code only
+sourceindex init opencode    # set up opencode only
+```
+
+It'll prompt for the key, set up some git hooks plus a subagent for the
+coding agent your repo already uses (a `CLAUDE.md` or `.claude/` means
+Claude Code; an `AGENTS.md` or `.opencode/` means opencode; neither means
+Claude Code), and scan your repo. A few minutes, once per repo. You can
+also name both at once (`sourceindex init claude opencode`), and it's safe
+to re-run init later to add the other agent.
+If you don't have a `CLAUDE.md` (Claude Code) or `AGENTS.md` (opencode)
+yet, run `/init` in your agent first.
+
+To replace a rotated or expired key later:
+
+```bash
+sourceindex api-key set
+```
+
+That prompts for the new key and persists it as your user-level default. New
+repos and repos without a local override will use it automatically. For
+automation, use
+`sourceindex api-key set --from-env` with `SOURCEINDEX_API_KEY` already set,
+pass `--api-key sk-si-...`, or pass the key positionally:
+`sourceindex api-key set sk-si-...`.
+
+To set a repo-local override:
+
+```bash
+sourceindex api-key set --local
+```
+
+Explicit environment variables still win, and a repo-local key overrides the
+global default.
+
+## Day-to-day
+
+Nothing. Open the repo in Claude Code or opencode as usual — it'll use
+SourceIndex on its own.
+
+## Feedback I'd love
+
+- Did `sourceindex init` finish cleanly?
+- Does the agent feel different? Even "I can't tell" is useful.
+- Anything in this doc that didn't land.
+
+More about the project lives at https://sourceindex.dev, and you can reach
+us any time at hi@sourceindex.dev.
+
+Thanks!

{sourceindex-0.1.0 → sourceindex-0.1.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "SourceIndex"
-version = "0.1.0"
+version = "0.1.2"
 description = "Codebase index for agentic coding"
 readme = "GETTING_STARTED.md"
 requires-python = ">=3.10"
@@ -32,9 +32,6 @@ dependencies = [
     "tree-sitter-typescript>=0.23",
 ]
 
-[project.optional-dependencies]
-mcp = ["mcp>=1.0.0,<2"]
-
 [project.scripts]
 sourceindex = "sourceindex.cli:main"
 

{sourceindex-0.1.0 → sourceindex-0.1.2}/sourceindex/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.0"
+__version__ = "0.1.2"
 
 # Surfaced at the top level so the CLI's argparse setup can use it for
 # --workers help/default without importing the build pipeline (which

{sourceindex-0.1.0 → sourceindex-0.1.2}/sourceindex/build/indexer.py

@@ -36,13 +36,20 @@ REALTIME_MAX_RETRIES = 5
 # file, but pays for itself at 2+.
 BATCH_MIN_FILES = 2
 RETRY_DELAY = 1.0
-# Best price/latency/quality balance found across DS-V4-Flash, Haiku 4.5,
-# Mistral-Small, Gemini-Flash-Lite, gpt-5-nano, gpt-5.4-nano (Opus-judged
-# description quality + AST-grounded recall + line-range accuracy).
+# The BUILD (indexing) default. Indexing is one LLM call per source file, so
+# this is the high-volume, cost-dominant model — hence the small cheap one.
+# Search keeps its own default (search.DEFAULT_SEARCH_MODEL): 1a/1b selection
+# quality is a different tradeoff from per-file summarization, and the backend
+# routes the two dimensions independently, so they default independently too.
+# Previous default: openrouter/openai/gpt-5.4-nano, picked as the best
+# price/latency/quality balance across DS-V4-Flash, Haiku 4.5, Mistral-Small,
+# Gemini-Flash-Lite and the gpt-5 nano family (Opus-judged description quality
+# + AST-grounded recall + line-range accuracy).
 # Reached via SOURCEINDEX_API_KEY: an sk-si bearer routes through the hosted
-# backend, an sk-or-/sk-ant-/etc. provider key auto-opts-out and hits the
-# provider directly. SOURCEINDEX_BACKEND_URL pins/disables the routing.
-DEFAULT_MODEL = "openrouter/openai/gpt-5.4-nano"
+# backend, which serves this model from AWS Bedrock (backend migration 0009);
+# an sk-or-/sk-ant-/etc. provider key auto-opts-out and hits the provider
+# directly. SOURCEINDEX_BACKEND_URL pins/disables the routing.
+DEFAULT_MODEL = "openrouter/google/gemma-4-e2b"
 BATCH_POLL_INTERVAL = 30
 BATCH_MAX_WAIT = 3600
 BATCH_MAX_TOKENS = 16384
@@ -102,7 +109,7 @@ def index_file(
     total_usage = LLMUsage()
     for attempt in range(2):
         try:
-            response, usage = call_llm(prompt, model=model, api_key=api_key)
+            response, usage = call_llm(prompt, model=model, api_key=api_key, pass_tag="build")
             total_usage.add(usage)
             if response:
                 oneliner, functions = parse_response(response)
@@ -305,6 +312,7 @@ def submit_and_wait_batch(
                     model=req["model"],
                     api_key=api_key,
                     max_tokens=req["max_tokens"],
+                    pass_tag="build",
                 )
                 attempt_usage.add(usage)
                 return cid, text or "", attempt_usage

{sourceindex-0.1.0 → sourceindex-0.1.2}/sourceindex/build/state.py

@@ -31,6 +31,10 @@ SUMMARY_CACHE_FILENAME = "_summary_cache.json"
 FILE_INDEX_FILENAME = "_file_index.json"
 BUILD_META_FILENAME = "build_meta.json"
 
+# Written as the details body when tier-2 indexing produced nothing; search
+# matches against this exact string to classify a file as un-indexed.
+INDEXING_FAILED_PLACEHOLDER = "(LLM indexing failed)"
+
 
 # ── Summary cache ──────────────────────────────────────────────────────────
 
@@ -117,10 +121,17 @@ def save_file_index(store, paths: dict[str, dict]) -> None:
 
 def write_details(store, rel_path: str, functions_text: str) -> None:
     """Write tier-2 function index to details/<rel_path>.md."""
-    body = functions_text.strip() or "(LLM indexing failed)"
+    body = functions_text.strip() or INDEXING_FAILED_PLACEHOLDER
     store.write_text(f"details/{rel_path}.md", f"# {rel_path}\n\n{body}\n")
 
 
+def details_failed(text: str) -> bool:
+    """True when a details file (as composed by write_details) holds only
+    the tier-2 failure placeholder."""
+    _, _, body = text.partition("\n\n")
+    return body.strip() == INDEXING_FAILED_PLACEHOLDER
+
+
 def details_exists(store, rel_path: str) -> bool:
     return store.exists(f"details/{rel_path}.md")