Class FileIndex

java.lang.Object

ai.redouble.demo.extract.FileIndex

public class FileIndex
extends Object

The rudimentary index: every extracted text with its vector, in memory, searched by cosine. What a deployment replaces with pgvector or any vector store; the shape a search needs is all that is here. Vectors are comparable only with vectors from the model that produced them, so the index records its model and refuses a query embedded by another. The index is also what the extraction leaves behind for the next workflow: every document's whole text, by path, which the pricing demo reads.

Since:

0.1 (2026-09-14)

Author:

Andrey Santrosyan

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static final record	FileIndex.Entry
static final record	FileIndex.Hit

Field Summary

Fields

Modifier and Type	Field	Description
static final int	INDEXED_CHARS	How much of a text is embedded: the head, which is what a search question is usually about.

Constructor Summary

Constructors

Constructor	Description
FileIndex()

Method Summary

Modifier and Type	Method	Description
void	clear()
String	getModelId()
void	put(String path, String text, Embedding embedding)
List<FileIndex.Hit>	search(Embedding query, int k)	The closest entries to a query vector, best first.
int	size()
Map<String,String>	texts()	Every indexed document's whole text by path, in path order: what a workflow over the extraction reads.

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

INDEXED_CHARS

public static final int INDEXED_CHARS

How much of a text is embedded: the head, which is what a search question is usually about. The whole text is kept.

See Also:

• Constant Field Values

Constructor Details

FileIndex

public FileIndex()

Method Details

put

public void put(String path,
 String text,
 Embedding embedding)

size

public int size()

texts

public Map<String,String> texts()

Every indexed document's whole text by path, in path order: what a workflow over the extraction reads.

getModelId

public String getModelId()

clear

public void clear()

search

public List<FileIndex.Hit> search(Embedding query,
 int k)

The closest entries to a query vector, best first.