Overview / RabbitHole

RabbitHole

Accessors

fileParsers

Get Signature

get fileParsers(): object

Get the file parsers

Returns

object

---

textSplitter

Get Signature

get textSplitter(): TextSplitter

Get the text splitter

Returns

TextSplitter

---

webParsers

Get Signature

get webParsers(): WebParser[]

Get the web parsers

Returns

WebParser[]

Methods

ingestContent()

ingestContent(stray, content, source, metadata?): Promise<void>

Ingests textual content into the memory.

Parameters

Parameter	Type	Default value	Description
stray	StrayCat	undefined	The StrayCat instance.
content	string | string[]	undefined	The textual content to ingest.
source	string	'unknown'	The source of the content (default: 'unknown').
metadata?	Record<string, any>	undefined	Additional metadata to store with the content.

Returns

Promise<void>

---

ingestFile()

ingestFile(stray, file, chunkSize?, chunkOverlap?, metadata?): Promise<void>

Ingests a file and processes its content.

Parameters

Parameter	Type	Description
stray	StrayCat	The StrayCat instance.
file	File	The file to ingest.
chunkSize?	number	The size of each chunk for splitting the content.
chunkOverlap?	number	The overlap between chunks.
metadata?	Record<string, any>	Additional metadata to store with the content.

Returns

Promise<void>

Throws

An error if the file type is not supported.

---

ingestMemory()

ingestMemory(json): Promise<void>

Upload memories to the declarative memory from a JSON file. When doing this, please, make sure the embedder used to export the memories is the same as the one used when uploading. The method also performs a check on the dimensionality of the embeddings (i.e. length of each vector).

Parameters

Parameter	Type	Description
json	MemoryJson | File	the json object containing the memories to be ingested.

Returns

Promise<void>

---

ingestPathOrURL()

ingestPathOrURL(stray, path, chunkSize?, chunkOverlap?, metadata?): Promise<void>

Ingests a path or URL and processes the content. If the input is a URL, it uses a web handler to load the content. If the input is a file system path, it reads the file and processes the content.

Parameters

Parameter	Type	Description
stray	StrayCat	The StrayCat instance.
path	string	The path or URL to ingest.
chunkSize?	number	The size of each chunk for splitting the content.
chunkOverlap?	number	The overlap between chunks.
metadata?	Record<string, any>	Additional metadata to store with the content.

Returns

Promise<void>

Throws

If the URL doesn't match any web handler or the path doesn't exist.

---

splitDocs()

splitDocs(stray, docs, chunkSize?, chunkOverlap?): Promise<Document[]>

Splits an array of texts into smaller chunks and creates documents. The method also executes the beforeSplitTexts and afterSplitTexts hooks.

Parameters

Parameter	Type	Description
stray	StrayCat	The StrayCat instance.
docs	Document[]	The array of documents to be split.
chunkSize?	number	The size of each chunk for splitting the content (default: 256).
chunkOverlap?	number	The overlap between chunks (default: 64).

Returns

Promise<Document[]>

An array of documents.

---

storeDocuments()

storeDocuments(stray, docs, source, metadata?): Promise<void>

Stores the given documents in memory. The method also executes the beforeStoreDocuments and beforeInsertInMemory hooks. It sends a websocket notification of the progress and when the reading process is completed

Parameters

Parameter	Type	Description
stray	StrayCat	The StrayCat instance.
docs	Document[]	An array of documents to store.
source	string	The source of the documents.
metadata?	Record<string, any>	Additional metadata to store with the content.

Returns

Promise<void>

---

getInstance()

static getInstance(): Promise<RabbitHole>

Get the Rabbit Hole instance

Returns

Promise<RabbitHole>

The Rabbit Hole class as a singleton