cocoindex 0.1.58__tar.gz → 0.1.59__tar.gz

cocoindex-0.1.59/.github/workflows/CI.yml

@@ -0,0 +1,31 @@
+# This file is autogenerated by maturin v1.8.1
+# To update, run
+#
+#    maturin generate-ci github
+#
+name: CI
+
+on:
+  pull_request:
+    branches: [main]
+    paths:
+      - src/**
+      - python/**
+      - "*.toml"
+      - ".github/workflows/*.yml"
+  push:
+    branches: [main]
+    paths:
+      - src/**
+      - python/**
+      - "*.toml"
+      - ".github/workflows/*.yml"
+  workflow_dispatch:
+
+permissions:
+  contents: read
+
+jobs:
+  test:
+    name: Run test
+    uses: ./.github/workflows/_test.yml

cocoindex-0.1.58/.github/workflows/CI.yml → cocoindex-0.1.59/.github/workflows/format.yml

@@ -3,7 +3,7 @@
 #
 #    maturin generate-ci github
 #
-name: CI
+name: format check
 
 on:
   pull_request:
@@ -11,15 +11,13 @@ on:
     paths:
       - src/**
       - python/**
-      - "*.toml"
-      - ".github/workflows/*.yml"
+      - examples/**
   push:
     branches: [main]
     paths:
       - src/**
       - python/**
-      - "*.toml"
-      - ".github/workflows/*.yml"
+      - examples/**
   workflow_dispatch:
 
 permissions:
@@ -51,8 +49,4 @@ jobs:
           pip install ruff
       - name: Check Python formatting
         run: |
-          ruff format --check .
-
-  test:
-    name: Run test
-    uses: ./.github/workflows/_test.yml
+          ruff format --check .

{cocoindex-0.1.58 → cocoindex-0.1.59}/Cargo.lock

@@ -1040,7 +1040,7 @@ dependencies = [
 
 [[package]]
 name = "cocoindex"
-version = "0.1.58"
+version = "0.1.59"
 dependencies = [
  "anyhow",
  "async-openai",

{cocoindex-0.1.58 → cocoindex-0.1.59}/Cargo.toml

@@ -2,7 +2,7 @@
 name = "cocoindex"
 # Version used for local development is always higher than others to take precedence.
 # Will be overridden for specific release versions.
-version = "0.1.58"
+version = "0.1.59"
 edition = "2024"
 rust-version = "1.86"
 

{cocoindex-0.1.58 → cocoindex-0.1.59}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cocoindex
-Version: 0.1.58
+Version: 0.1.59
 Requires-Dist: click>=8.1.8
 Requires-Dist: rich>=14.0.0
 Requires-Dist: python-dotenv>=1.1.0

{cocoindex-0.1.58 → cocoindex-0.1.59}/docs/docs/ai/llm.mdx

@@ -26,6 +26,7 @@ We support the following types of LLM APIs:
 | [Voyage](#voyage) | `LlmApiType.VOYAGE` | ❌ | ✅ |
 | [LiteLLM](#litellm) | `LlmApiType.LITE_LLM` | ✅ | ❌ |
 | [OpenRouter](#openrouter) | `LlmApiType.OPEN_ROUTER` | ✅ | ❌ |
+| [vLLM](#vllm) | `LlmApiType.VLLM` | ✅ | ❌ |
 
 ## LLM Tasks
 
@@ -307,3 +308,34 @@ cocoindex.LlmSpec(
 </Tabs>
 
 You can find the full list of models supported by OpenRouter [here](https://openrouter.ai/models).
+
+### vLLM
+
+Install vLLM:
+
+```bash
+pip install vllm
+```
+
+Run vLLM Server
+
+```bash
+vllm serve deepseek-ai/deepseek-coder-1.3b-instruct
+```
+
+
+A spec for vLLM looks like this:
+
+<Tabs>
+<TabItem value="python" label="Python" default>
+
+```python
+cocoindex.LlmSpec(
+    api_type=cocoindex.LlmApiType.VLLM,
+    model="deepseek-ai/deepseek-coder-1.3b-instruct",
+    address="http://127.0.0.1:8000/v1",
+)
+```
+
+</TabItem>
+</Tabs>

{cocoindex-0.1.58 → cocoindex-0.1.59}/docs/docs/core/data_types.mdx

@@ -38,6 +38,7 @@ This is the list of all basic types supported by CocoIndex:
 | TimeDelta | A duration of time | `datetime.timedelta` | `datetime.timedelta` |
 | Json | | `cocoindex.Json` | Any data convertible to JSON by `json` package | 
 | Vector[*T*, *Dim*?] | *T* can be a basic type or a numeric type. *Dim* is a positive integer and optional. | `cocoindex.Vector[T]` or `cocoindex.Vector[T, Dim]` | `numpy.typing.NDArray[T]` or `list[T]` |
+| Union[*T1*, *T2*, ...] | *T1*, *T2*, ... are any basic types | `T1 | T2 | ...` | `T1 | T2 | ...` |
 
 Values of all data types can be represented by values in Python's native types (as described under the Native Python Type column).
 However, the underlying execution engine has finer distinctions for some types, specifically:

cocoindex-0.1.59/examples/patient_intake_extraction/.env.example

@@ -0,0 +1,4 @@
+# Postgres database address for cocoindex
+COCOINDEX_DATABASE_URL=postgres://cocoindex:cocoindex@localhost/cocoindex
+
+OPENAI_API_KEY=

cocoindex-0.1.59/examples/patient_intake_extraction/README.md

@@ -0,0 +1,58 @@
+# Extract structured data from patient intake forms with LLM
+[![GitHub](https://img.shields.io/github/stars/cocoindex-io/cocoindex?color=5B5BD6)](https://github.com/cocoindex-io/cocoindex)
+
+
+This repo shows how to use OpenAI API to extract structured data from patient intake forms with different formats, like PDF, Docx, etc. from Google Drive.
+
+We appreciate a star ⭐ at [CocoIndex Github](https://github.com/cocoindex-io/cocoindex) if this is helpful.
+
+![Structured Data From Patient Intake Forms](https://github.com/user-attachments/assets/1f6afb69-d26d-4a08-8774-13982d6aec1e)
+
+
+## Tutorials
+- Step by step tutorial - Check out the [blog](https://cocoindex.io/blogs/patient-intake-form-extraction-with-llm).
+- Video tutorial - [Youtube](https://youtu.be/_mjlwVtnBn0?si=cpH-4kkOAYm2HhK6).
+
+## Prerequisite
+1. [Install Postgres](https://cocoindex.io/docs/getting_started/installation#-install-postgres) if you don't have one.
+
+2. Install CocoIndex
+   ```bash
+   pip install -U cocoindex
+   ```
+
+3. Install MarkItDown
+   ```bash
+   pip install 'markitdown[all]'
+   ```
+4. Create a `.env` file from `.env.example`, and fill `OPENAI_API_KEY`.
+
+## Run
+
+Setup index:
+
+```bash
+cocoindex setup main.py
+```
+
+Update index:
+
+```bash
+cocoindex update main.py
+```
+
+Run query:
+
+```bash
+python main.py
+```
+
+Run with CocoInsight:
+```bash
+cocoindex server -ci main.py
+```
+<img width="1405" alt="Screenshot 2025-07-02 at 11 59 24 AM" src="https://github.com/user-attachments/assets/6f5154cd-8a53-4baa-b914-cd60ffecf3d4" />
+
+
+
+View results at https://cocoindex.io/cocoinsight

cocoindex-0.1.59/examples/patient_intake_extraction/data/README.md

@@ -0,0 +1,4 @@
+## Note:
+Example files here are purely artificial and not real, for testing purposes only.
+Please do not use these examples for any other purpose.
+

cocoindex-0.1.59/examples/patient_intake_extraction/main.py

@@ -0,0 +1,148 @@
+import datetime
+import tempfile
+import dataclasses
+import os
+
+from markitdown import MarkItDown
+from openai import OpenAI
+
+import cocoindex
+
+
+@dataclasses.dataclass
+class Contact:
+    name: str
+    phone: str
+    relationship: str
+
+
+@dataclasses.dataclass
+class Address:
+    street: str
+    city: str
+    state: str
+    zip_code: str
+
+
+@dataclasses.dataclass
+class Pharmacy:
+    name: str
+    phone: str
+    address: Address
+
+
+@dataclasses.dataclass
+class Insurance:
+    provider: str
+    policy_number: str
+    group_number: str | None
+    policyholder_name: str
+    relationship_to_patient: str
+
+
+@dataclasses.dataclass
+class Condition:
+    name: str
+    diagnosed: bool
+
+
+@dataclasses.dataclass
+class Medication:
+    name: str
+    dosage: str
+
+
+@dataclasses.dataclass
+class Allergy:
+    name: str
+
+
+@dataclasses.dataclass
+class Surgery:
+    name: str
+    date: str
+
+
+@dataclasses.dataclass
+class Patient:
+    name: str
+    dob: datetime.date
+    gender: str
+    address: Address
+    phone: str
+    email: str
+    preferred_contact_method: str
+    emergency_contact: Contact
+    insurance: Insurance | None
+    reason_for_visit: str
+    symptoms_duration: str
+    past_conditions: list[Condition]
+    current_medications: list[Medication]
+    allergies: list[Allergy]
+    surgeries: list[Surgery]
+    occupation: str | None
+    pharmacy: Pharmacy | None
+    consent_given: bool
+    consent_date: datetime.date | None
+
+
+class ToMarkdown(cocoindex.op.FunctionSpec):
+    """Convert a document to markdown."""
+
+
+@cocoindex.op.executor_class(gpu=True, cache=True, behavior_version=1)
+class ToMarkdownExecutor:
+    """Executor for ToMarkdown."""
+
+    spec: ToMarkdown
+    _converter: MarkItDown
+
+    def prepare(self):
+        client = OpenAI()
+        self._converter = MarkItDown(llm_client=client, llm_model="gpt-4o")
+
+    def __call__(self, content: bytes, filename: str) -> str:
+        suffix = os.path.splitext(filename)[1]
+        with tempfile.NamedTemporaryFile(delete=True, suffix=suffix) as temp_file:
+            temp_file.write(content)
+            temp_file.flush()
+            text = self._converter.convert(temp_file.name).text_content
+            return text
+
+
+@cocoindex.flow_def(name="PatientIntakeExtraction")
+def patient_intake_extraction_flow(
+    flow_builder: cocoindex.FlowBuilder, data_scope: cocoindex.DataScope
+):
+    """
+    Define a flow that extracts patient information from intake forms.
+    """
+    data_scope["documents"] = flow_builder.add_source(
+        cocoindex.sources.LocalFile(path="data/patient_forms", binary=True)
+    )
+
+    patients_index = data_scope.add_collector()
+
+    with data_scope["documents"].row() as doc:
+        doc["markdown"] = doc["content"].transform(
+            ToMarkdown(), filename=doc["filename"]
+        )
+        doc["patient_info"] = doc["markdown"].transform(
+            cocoindex.functions.ExtractByLlm(
+                llm_spec=cocoindex.LlmSpec(
+                    api_type=cocoindex.LlmApiType.OPENAI, model="gpt-4o"
+                ),
+                output_type=Patient,
+                instruction="Please extract patient information from the intake form.",
+            )
+        )
+        patients_index.collect(
+            filename=doc["filename"],
+            patient_info=doc["patient_info"],
+        )
+
+    patients_index.export(
+        "patients",
+        cocoindex.storages.Postgres(table_name="patients_info"),
+        primary_key_fields=["filename"],
+    )

cocoindex-0.1.59/examples/patient_intake_extraction/pyproject.toml

@@ -0,0 +1,11 @@
+[project]
+name = "patient-intake-extraction"
+version = "0.1.0"
+description = "Extract structured information from patient intake forms using LLM."
+requires-python = ">=3.10"
+dependencies = [
+    "cocoindex>=0.1.45",
+    "python-dotenv>=1.0.1",
+    "markitdown>=0.1.2",
+    "openai>=1.68.2"
+]

{cocoindex-0.1.58 → cocoindex-0.1.59}/python/cocoindex/convert.py

@@ -5,7 +5,6 @@ Utilities to convert between Python and engine values.
 import dataclasses
 import datetime
 import inspect
-import uuid
 from enum import Enum
 from typing import Any, Callable, Mapping, get_origin
 
@@ -14,7 +13,6 @@ import numpy as np
 from .typing import (
     KEY_FIELD_NAME,
     TABLE_TYPES,
-    AnalyzedTypeInfo,
     DtypeRegistry,
     analyze_type_info,
     encode_enriched_type,
@@ -74,23 +72,16 @@ def make_engine_value_decoder(
     Returns:
         A decoder from an engine value to a Python value.
     """
-
     src_type_kind = src_type["kind"]
 
-    dst_type_info: AnalyzedTypeInfo | None = None
-    if (
-        dst_annotation is not None
-        and dst_annotation is not inspect.Parameter.empty
-        and dst_annotation is not Any
-    ):
-        dst_type_info = analyze_type_info(dst_annotation)
-        if not _is_type_kind_convertible_to(src_type_kind, dst_type_info.kind):
-            raise ValueError(
-                f"Type mismatch for `{''.join(field_path)}`: "
-                f"passed in {src_type_kind}, declared {dst_annotation} ({dst_type_info.kind})"
-            )
-
-    if dst_type_info is None:
+    dst_is_any = (
+        dst_annotation is None
+        or dst_annotation is inspect.Parameter.empty
+        or dst_annotation is Any
+    )
+    if dst_is_any:
+        if src_type_kind == "Union":
+            return lambda value: value[1]
         if src_type_kind == "Struct" or src_type_kind in TABLE_TYPES:
             raise ValueError(
                 f"Missing type annotation for `{''.join(field_path)}`."
@@ -98,6 +89,41 @@ def make_engine_value_decoder(
             )
         return lambda value: value
 
+    dst_type_info = analyze_type_info(dst_annotation)
+
+    if src_type_kind == "Union":
+        dst_type_variants = (
+            dst_type_info.union_variant_types
+            if dst_type_info.union_variant_types is not None
+            else [dst_annotation]
+        )
+        src_type_variants = src_type["types"]
+        decoders = []
+        for i, src_type_variant in enumerate(src_type_variants):
+            src_field_path = field_path + [f"[{i}]"]
+            decoder = None
+            for dst_type_variant in dst_type_variants:
+                try:
+                    decoder = make_engine_value_decoder(
+                        src_field_path, src_type_variant, dst_type_variant
+                    )
+                    break
+                except ValueError:
+                    pass
+            if decoder is None:
+                raise ValueError(
+                    f"Type mismatch for `{''.join(field_path)}`: "
+                    f"cannot find matched target type for source type variant {src_type_variant}"
+                )
+            decoders.append(decoder)
+        return lambda value: decoders[value[0]](value[1])
+
+    if not _is_type_kind_convertible_to(src_type_kind, dst_type_info.kind):
+        raise ValueError(
+            f"Type mismatch for `{''.join(field_path)}`: "
+            f"passed in {src_type_kind}, declared {dst_annotation} ({dst_type_info.kind})"
+        )
+
     if dst_type_info.kind in ("Float32", "Float64", "Int64"):
         dst_core_type = dst_type_info.core_type
 
@@ -196,9 +222,6 @@ def make_engine_value_decoder(
         field_path.pop()
         return decode
 
-    if src_type_kind == "Union":
-        return lambda value: value[1]
-
     return lambda value: value
 
 

{cocoindex-0.1.58 → cocoindex-0.1.59}/python/cocoindex/llm.py

@@ -12,6 +12,7 @@ class LlmApiType(Enum):
     LITE_LLM = "LiteLlm"
     OPEN_ROUTER = "OpenRouter"
     VOYAGE = "Voyage"
+    VLLM = "Vllm"
 
 
 @dataclass

{cocoindex-0.1.58 → cocoindex-0.1.59}/python/cocoindex/tests/test_convert.py

@@ -104,7 +104,7 @@ def validate_full_roundtrip(
     )
     decoder = make_engine_value_decoder([], encoded_output_type, value_type)
     decoded_value = decoder(value_from_engine)
-    assert eq(decoded_value, value)
+    assert eq(decoded_value, value), f"{decoded_value} != {value}"
 
     if other_decoded_values is not None:
         for other_value, other_type in other_decoded_values:
@@ -613,6 +613,18 @@ def test_roundtrip_union_timedelta() -> None:
     validate_full_roundtrip(value, t)
 
 
+def test_roundtrip_vector_of_union() -> None:
+    t = list[str | int]
+    value = ["a", 1]
+    validate_full_roundtrip(value, t)
+
+
+def test_roundtrip_union_with_vector() -> None:
+    t = NDArray[np.float32] | str
+    value = np.array([1.0, 2.0, 3.0], dtype=np.float32)
+    validate_full_roundtrip(value, t, ([1.0, 2.0, 3.0], list[float] | str))
+
+
 def test_roundtrip_ltable() -> None:
     t = list[Order]
     value = [Order("O1", "item1", 10.0), Order("O2", "item2", 20.0)]

{cocoindex-0.1.58 → cocoindex-0.1.59}/src/llm/mod.rs

@@ -13,6 +13,7 @@ pub enum LlmApiType {
     LiteLlm,
     OpenRouter,
     Voyage,
+    Vllm,
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
@@ -81,6 +82,7 @@ mod litellm;
 mod ollama;
 mod openai;
 mod openrouter;
+mod vllm;
 mod voyage;
 
 pub async fn new_llm_generation_client(
@@ -108,6 +110,9 @@ pub async fn new_llm_generation_client(
         LlmApiType::Voyage => {
             api_bail!("Voyage is not supported for generation")
         }
+        LlmApiType::Vllm => {
+            Box::new(vllm::Client::new_vllm(address).await?) as Box<dyn LlmGenerationClient>
+        }
     };
     Ok(client)
 }
@@ -129,6 +134,7 @@ pub fn new_llm_embedding_client(
         LlmApiType::Ollama
         | LlmApiType::OpenRouter
         | LlmApiType::LiteLlm
+        | LlmApiType::Vllm
         | LlmApiType::Anthropic => {
             api_bail!("Embedding is not supported for API type {:?}", api_type)
         }

cocoindex-0.1.59/src/llm/vllm.rs

@@ -0,0 +1,16 @@
+use async_openai::Client as OpenAIClient;
+use async_openai::config::OpenAIConfig;
+
+pub use super::openai::Client;
+
+impl Client {
+    pub async fn new_vllm(address: Option<String>) -> anyhow::Result<Self> {
+        let address = address.unwrap_or_else(|| "http://127.0.0.1:8000/v1".to_string());
+        let api_key = std::env::var("VLLM_API_KEY").ok();
+        let mut config = OpenAIConfig::new().with_api_base(address);
+        if let Some(api_key) = api_key {
+            config = config.with_api_key(api_key);
+        }
+        Ok(Client::from_parts(OpenAIClient::with_config(config)))
+    }
+}