Merge pull request #3073 from skylares/fireflies-dev

Fireflies connector

backend/danswer/configs/constants.py

@@ -126,6 +126,7 @@ class DocumentSource(str, Enum):
     XENFORO = "xenforo"
     NOT_APPLICABLE = "not_applicable"
     FRESHDESK = "freshdesk"
+    FIREFLIES = "fireflies"
 
 
 DocumentSourceRequiringTenantContext: list[DocumentSource] = [DocumentSource.FILE]

backend/danswer/connectors/factory.py

@@ -16,6 +16,7 @@ from danswer.connectors.discourse.connector import DiscourseConnector
 from danswer.connectors.document360.connector import Document360Connector
 from danswer.connectors.dropbox.connector import DropboxConnector
 from danswer.connectors.file.connector import LocalFileConnector
+from danswer.connectors.fireflies.connector import FirefliesConnector
 from danswer.connectors.freshdesk.connector import FreshdeskConnector
 from danswer.connectors.github.connector import GithubConnector
 from danswer.connectors.gitlab.connector import GitlabConnector
@@ -101,6 +102,7 @@ def identify_connector_class(
         DocumentSource.OCI_STORAGE: BlobStorageConnector,
         DocumentSource.XENFORO: XenforoConnector,
         DocumentSource.FRESHDESK: FreshdeskConnector,
+        DocumentSource.FIREFLIES: FirefliesConnector,
     }
     connector_by_source = connector_map.get(source, {})
 

backend/danswer/connectors/fireflies/connector.py

@@ -0,0 +1,182 @@
+from collections.abc import Iterator
+from datetime import datetime
+from datetime import timezone
+from typing import List
+
+import requests
+
+from danswer.configs.app_configs import INDEX_BATCH_SIZE
+from danswer.configs.constants import DocumentSource
+from danswer.connectors.interfaces import GenerateDocumentsOutput
+from danswer.connectors.interfaces import LoadConnector
+from danswer.connectors.interfaces import PollConnector
+from danswer.connectors.interfaces import SecondsSinceUnixEpoch
+from danswer.connectors.models import BasicExpertInfo
+from danswer.connectors.models import ConnectorMissingCredentialError
+from danswer.connectors.models import Document
+from danswer.connectors.models import Section
+from danswer.utils.logger import setup_logger
+
+logger = setup_logger()
+
+_FIREFLIES_ID_PREFIX = "FIREFLIES_"
+
+_FIREFLIES_API_URL = "https://api.fireflies.ai/graphql"
+
+_FIREFLIES_TRANSCRIPT_QUERY_SIZE = 50  # Max page size is 50
+
+_FIREFLIES_API_QUERY = """
+    query Transcripts($fromDate: DateTime, $toDate: DateTime, $limit: Int!, $skip: Int!) {
+        transcripts(fromDate: $fromDate, toDate: $toDate, limit: $limit, skip: $skip) {
+            id
+            title
+            host_email
+            participants
+            date
+            transcript_url
+            sentences {
+                text
+                speaker_name
+            }
+        }
+    }
+"""
+
+
+def _create_doc_from_transcript(transcript: dict) -> Document | None:
+    meeting_text = ""
+    sentences = transcript.get("sentences", [])
+    if sentences:
+        for sentence in sentences:
+            meeting_text += sentence.get("speaker_name") or "Unknown Speaker"
+            meeting_text += ": " + sentence.get("text", "") + "\n\n"
+    else:
+        return None
+
+    meeting_link = transcript["transcript_url"]
+
+    fireflies_id = _FIREFLIES_ID_PREFIX + transcript["id"]
+
+    meeting_title = transcript["title"] or "No Title"
+
+    meeting_date_unix = transcript["date"]
+    meeting_date = datetime.fromtimestamp(meeting_date_unix / 1000, tz=timezone.utc)
+
+    meeting_host_email = transcript["host_email"]
+    host_email_user_info = [BasicExpertInfo(email=meeting_host_email)]
+
+    meeting_participants_email_list = []
+    for participant in transcript.get("participants", []):
+        if participant != meeting_host_email and participant:
+            meeting_participants_email_list.append(BasicExpertInfo(email=participant))
+
+    return Document(
+        id=fireflies_id,
+        sections=[
+            Section(
+                link=meeting_link,
+                text=meeting_text,
+            )
+        ],
+        source=DocumentSource.FIREFLIES,
+        semantic_identifier=meeting_title,
+        metadata={},
+        doc_updated_at=meeting_date,
+        primary_owners=host_email_user_info,
+        secondary_owners=meeting_participants_email_list,
+    )
+
+
+class FirefliesConnector(PollConnector, LoadConnector):
+    def __init__(self, batch_size: int = INDEX_BATCH_SIZE) -> None:
+        self.batch_size = batch_size
+
+    def load_credentials(self, credentials: dict[str, str]) -> None:
+        api_key = credentials.get("fireflies_api_key")
+
+        if not isinstance(api_key, str):
+            raise ConnectorMissingCredentialError(
+                "The Fireflies API key must be a string"
+            )
+
+        self.api_key = api_key
+
+        return None
+
+    def _fetch_transcripts(
+        self, start_datetime: str | None = None, end_datetime: str | None = None
+    ) -> Iterator[List[dict]]:
+        if self.api_key is None:
+            raise ConnectorMissingCredentialError("Missing API key")
+
+        headers = {
+            "Content-Type": "application/json",
+            "Authorization": "Bearer " + self.api_key,
+        }
+
+        skip = 0
+        variables: dict[str, int | str] = {
+            "limit": _FIREFLIES_TRANSCRIPT_QUERY_SIZE,
+        }
+
+        if start_datetime:
+            variables["fromDate"] = start_datetime
+        if end_datetime:
+            variables["toDate"] = end_datetime
+
+        while True:
+            variables["skip"] = skip
+            response = requests.post(
+                _FIREFLIES_API_URL,
+                headers=headers,
+                json={"query": _FIREFLIES_API_QUERY, "variables": variables},
+            )
+
+            response.raise_for_status()
+
+            if response.status_code == 204:
+                break
+
+            recieved_transcripts = response.json()
+            parsed_transcripts = recieved_transcripts.get("data", {}).get(
+                "transcripts", []
+            )
+
+            yield parsed_transcripts
+
+            if len(parsed_transcripts) < _FIREFLIES_TRANSCRIPT_QUERY_SIZE:
+                break
+
+            skip += _FIREFLIES_TRANSCRIPT_QUERY_SIZE
+
+    def _process_transcripts(
+        self, start: str | None = None, end: str | None = None
+    ) -> GenerateDocumentsOutput:
+        doc_batch: List[Document] = []
+
+        for transcript_batch in self._fetch_transcripts(start, end):
+            for transcript in transcript_batch:
+                if doc := _create_doc_from_transcript(transcript):
+                    doc_batch.append(doc)
+
+                if len(doc_batch) >= self.batch_size:
+                    yield doc_batch
+                    doc_batch = []
+
+        if doc_batch:
+            yield doc_batch
+
+    def load_from_state(self) -> GenerateDocumentsOutput:
+        return self._process_transcripts()
+
+    def poll_source(
+        self, start_unixtime: SecondsSinceUnixEpoch, end_unixtime: SecondsSinceUnixEpoch
+    ) -> GenerateDocumentsOutput:
+        start_datetime = datetime.fromtimestamp(
+            start_unixtime, tz=timezone.utc
+        ).strftime("%Y-%m-%dT%H:%M:%S.000Z")
+        end_datetime = datetime.fromtimestamp(end_unixtime, tz=timezone.utc).strftime(
+            "%Y-%m-%dT%H:%M:%S.000Z"
+        )
+
+        yield from self._process_transcripts(start_datetime, end_datetime)

web/src/components/icons/icons.tsx

@@ -67,6 +67,7 @@ import s3Icon from "../../../public/S3.png";
 import r2Icon from "../../../public/r2.png";
 import salesforceIcon from "../../../public/Salesforce.png";
 import freshdeskIcon from "../../../public/Freshdesk.png";
+import firefliesIcon from "../../../public/Fireflies.png";
 
 import sharepointIcon from "../../../public/Sharepoint.png";
 import teamsIcon from "../../../public/Teams.png";
@@ -1293,6 +1294,13 @@ export const FreshdeskIcon = ({
   <LogoIcon size={size} className={className} src={freshdeskIcon} />
 );
 
+export const FirefliesIcon = ({
+  size = 16,
+  className = defaultTailwindCSS,
+}: IconProps) => (
+  <LogoIcon size={size} className={className} src={firefliesIcon} />
+);
+
 /* 
 EE Icons
 */

web/src/lib/connectors/connectors.tsx

@@ -950,6 +950,11 @@ For example, specifying .*-support.* as a "channel" will cause the connector to
     values: [],
     advanced_values: [],
   },
+  fireflies: {
+    description: "Configure Fireflies connector",
+    values: [],
+    advanced_values: [],
+  },
 };
 export function createConnectorInitialValues(
   connector: ConfigurableSources
@@ -1209,6 +1214,8 @@ export interface AsanaConfig {
 
 export interface FreshdeskConfig {}
 
+export interface FirefliesConfig {}
+
 export interface MediaWikiConfig extends MediaWikiBaseConfig {
   hostname: string;
 }

web/src/lib/connectors/credentials.ts

@@ -188,6 +188,10 @@ export interface FreshdeskCredentialJson {
   freshdesk_api_key: string;
 }
 
+export interface FirefliesCredentialJson {
+  fireflies_api_key: string;
+}
+
 export interface MediaWikiCredentialJson {}
 export interface WikipediaCredentialJson extends MediaWikiCredentialJson {}
 
@@ -291,6 +295,9 @@ export const credentialTemplates: Record<ValidSources, any> = {
     freshdesk_password: "",
     freshdesk_api_key: "",
   } as FreshdeskCredentialJson,
+  fireflies: {
+    fireflies_api_key: "",
+  } as FirefliesCredentialJson,
   xenforo: null,
   google_sites: null,
   file: null,
@@ -427,7 +434,11 @@ export const credentialDisplayNames: Record<string, string> = {
   freshdesk_domain: "Freshdesk Domain",
   freshdesk_password: "Freshdesk Password",
   freshdesk_api_key: "Freshdesk API Key",
+
+  // Fireflies
+  fireflies_api_key: "Fireflies API Key",
 };
+
 export function getDisplayNameForCredentialKey(key: string): string {
   return credentialDisplayNames[key] || key;
 }

web/src/lib/sources.ts

@@ -37,6 +37,7 @@ import {
   ColorSlackIcon,
   XenforoIcon,
   FreshdeskIcon,
+  FirefliesIcon,
 } from "@/components/icons/icons";
 import { ValidSources } from "./types";
 import {
@@ -289,6 +290,12 @@ const SOURCE_METADATA_MAP: SourceMap = {
     category: SourceCategory.CustomerSupport,
     docs: "https://docs.danswer.dev/connectors/freshdesk",
   },
+  fireflies: {
+    icon: FirefliesIcon,
+    displayName: "Fireflies",
+    category: SourceCategory.Other,
+    docs: "https://docs.danswer.dev/connectors/fireflies",
+  },
   // currently used for the Internet Search tool docs, which is why
   // a globe is used
   not_applicable: {

web/src/lib/types.ts

@@ -266,6 +266,7 @@ const validSources = [
   "not_applicable",
   "ingestion_api",
   "freshdesk",
+  "fireflies",
 ] as const;
 
 export type ValidSources = (typeof validSources)[number];