Implement freshdesk frontend

backend/danswer/configs/constants.py

@@ -128,6 +128,7 @@ class DocumentSource(str, Enum):
     OCI_STORAGE = "oci_storage"
     XENFORO = "xenforo"
     NOT_APPLICABLE = "not_applicable"
+    FRESHDESK = "freshdesk"
 
 
 DocumentSourceRequiringTenantContext: list[DocumentSource] = [DocumentSource.FILE]

backend/danswer/connectors/factory.py

@@ -46,6 +46,7 @@ from danswer.connectors.wikipedia.connector import WikipediaConnector
 from danswer.connectors.xenforo.connector import XenforoConnector
 from danswer.connectors.zendesk.connector import ZendeskConnector
 from danswer.connectors.zulip.connector import ZulipConnector
+from danswer.connectors.freshdesk.connector import FreshdeskConnector
 from danswer.db.credentials import backend_update_credential_json
 from danswer.db.models import Credential
 
@@ -101,6 +102,7 @@ def identify_connector_class(
         DocumentSource.GOOGLE_CLOUD_STORAGE: BlobStorageConnector,
         DocumentSource.OCI_STORAGE: BlobStorageConnector,
         DocumentSource.XENFORO: XenforoConnector,
+        DocumentSource.FRESHDESK: FreshdeskConnector,
     }
     connector_by_source = connector_map.get(source, {})
 

backend/danswer/connectors/freshdesk/connector.py

@@ -0,0 +1,116 @@
+import requests
+import json
+from datetime import datetime, timezone
+from typing import Any, List, Optional
+from bs4 import BeautifulSoup  # Add this import for HTML parsing
+from danswer.configs.app_configs import INDEX_BATCH_SIZE
+from danswer.configs.constants import DocumentSource
+from danswer.connectors.interfaces import GenerateDocumentsOutput, PollConnector
+from danswer.connectors.models import ConnectorMissingCredentialError, Document, Section
+from danswer.utils.logger import setup_logger
+
+logger = setup_logger()
+
+
+class FreshdeskConnector(PollConnector):
+    def __init__(self, api_key: str, domain: str, password: str, batch_size: int = INDEX_BATCH_SIZE) -> None:
+        self.api_key = api_key
+        self.domain = domain
+        self.password = password
+        self.batch_size = batch_size
+
+    def ticket_link(self, tid: int) -> str:
+        return f"https://{self.domain}.freshdesk.com/helpdesk/tickets/{tid}"
+
+    def build_doc_sections_from_ticket(self, ticket: dict) -> List[Section]:
+        # Use list comprehension for building sections
+        return [
+            Section(
+                link=self.ticket_link(int(ticket["id"])),
+                text=json.dumps({
+                    key: value
+                    for key, value in ticket.items()
+                    if isinstance(value, str)
+                }, default=str),
+            )
+        ]
+
+    def strip_html_tags(self, html: str) -> str:
+        soup = BeautifulSoup(html, 'html.parser')
+        return soup.get_text()
+
+    def load_credentials(self, credentials: dict[str, Any]) -> Optional[dict[str, Any]]:
+        logger.info("Loading credentials")
+        self.api_key = credentials.get("freshdesk_api_key")
+        self.domain = credentials.get("freshdesk_domain")
+        self.password = credentials.get("freshdesk_password")
+        return None
+
+    def _process_tickets(self, start: datetime, end: datetime) -> GenerateDocumentsOutput:
+        logger.info("Processing tickets")
+        if any([self.api_key, self.domain, self.password]) is None:
+            raise ConnectorMissingCredentialError("freshdesk")
+
+        freshdesk_url = f"https://{self.domain}.freshdesk.com/api/v2/tickets?include=description"
+        response = requests.get(freshdesk_url, auth=(self.api_key, self.password))
+        response.raise_for_status()  # raises exception when not a 2xx response
+
+        if response.status_code!= 204:
+            tickets = json.loads(response.content)
+            logger.info(f"Fetched {len(tickets)} tickets from Freshdesk API")
+            doc_batch: List[Document] = []
+
+            for ticket in tickets:
+                # Convert the "created_at", "updated_at", and "due_by" values to ISO 8601 strings
+                for date_field in ["created_at", "updated_at", "due_by"]:
+                    ticket[date_field] = datetime.fromisoformat(ticket[date_field]).strftime("%Y-%m-%d %H:%M:%S")
+
+                # Convert all other values to strings
+                ticket = {
+                    key: str(value) if not isinstance(value, str) else value
+                    for key, value in ticket.items()
+                }
+
+                # Checking for overdue tickets
+                today = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
+                ticket["overdue"] = "true" if today > ticket["due_by"] else "false"
+
+                # Mapping the status field values
+                status_mapping = {2: "open", 3: "pending", 4: "resolved", 5: "closed"}
+                ticket["status"] = status_mapping.get(ticket["status"], str(ticket["status"]))
+
+                # Stripping HTML tags from the description field
+                ticket["description"] = self.strip_html_tags(ticket["description"])
+
+                # Remove extra white spaces from the description field
+                ticket["description"] = " ".join(ticket["description"].split())
+
+                # Use list comprehension for building sections
+                sections = self.build_doc_sections_from_ticket(ticket)
+
+                created_at = datetime.fromisoformat(ticket["created_at"])
+                today = datetime.now()
+                if (today - created_at).days / 30.4375 <= 2:
+                    doc = Document(
+                        id=ticket["id"],
+                        sections=sections,
+                        source=DocumentSource.FRESHDESK,
+                        semantic_identifier=ticket["subject"],
+                        metadata={
+                            key: value
+                            for key, value in ticket.items()
+                            if isinstance(value, str) and key not in ["description", "description_text"]
+                        },
+                    )
+
+                    doc_batch.append(doc)
+
+                    if len(doc_batch) >= self.batch_size:
+                        yield doc_batch
+                        doc_batch = []
+
+            if doc_batch:
+                yield doc_batch
+
+    def poll_source(self, start: datetime, end: datetime) -> GenerateDocumentsOutput:
+        yield from self._process_tickets(start, end)

backend/danswer/db/models.py

@@ -3,7 +3,7 @@ import json
 from enum import Enum as PyEnum
 from typing import Any
 from typing import Literal
-from typing import NotRequired
+from typing_extensions import NotRequired
 from typing import Optional
 from uuid import uuid4
 from typing_extensions import TypedDict  # noreorder

backend/danswer/file_store/models.py

@@ -1,6 +1,6 @@
 import base64
 from enum import Enum
-from typing import NotRequired
+from typing_extensions import NotRequired
 from typing_extensions import TypedDict  # noreorder
 
 from pydantic import BaseModel

backend/model_server/main.py

@@ -28,7 +28,7 @@ os.environ["TOKENIZERS_PARALLELISM"] = "false"
 os.environ["HF_HUB_DISABLE_TELEMETRY"] = "1"
 
 HF_CACHE_PATH = Path("/root/.cache/huggingface/")
-TEMP_HF_CACHE_PATH = Path("/root/.cache/temp_huggingface/")
+TEMP_HF_CACHE_PATH = Path.home() / ".cache" / "temp_huggingface"
 
 transformer_logging.set_verbosity_error()
 

web/src/app/api/[...path]/route.ts

@@ -75,9 +75,13 @@ async function handleRequest(request: NextRequest, path: string[]) {
       backendUrl.searchParams.append(key, value);
     });
 
+    // Create a new headers object, omitting the 'connection' header
+    const headers = new Headers(request.headers);
+    headers.delete('connection');
+
     const response = await fetch(backendUrl, {
       method: request.method,
-      headers: request.headers,
+      headers: headers,
       body: request.body,
       signal: request.signal,
       // @ts-ignore

web/src/components/icons/icons.tsx

@@ -75,6 +75,7 @@ import slackIcon from "../../../public/Slack.png";
 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 sharepointIcon from "../../../public/Sharepoint.png";
 import teamsIcon from "../../../public/Teams.png";
@@ -1301,6 +1302,13 @@ export const AsanaIcon = ({
   className = defaultTailwindCSS,
 }: IconProps) => <LogoIcon size={size} className={className} src={asanaIcon} />;
 
+export const FreshdeskIcon = ({
+  size = 16,
+  className = defaultTailwindCSS,
+}: IconProps) => (
+  <LogoIcon size={size} className={className} src={freshdeskIcon} />
+);
+
 /* 
 EE Icons
 */

web/src/lib/connectors/connectors.tsx

@@ -922,6 +922,12 @@ For example, specifying .*-support.* as a "channel" will cause the connector to
     ],
     advanced_values: [],
   },
+  freshdesk: {
+    description: "Configure Freshdesk connector",
+    values: [],
+    advanced_values: [],
+  },
+  
 };
 export function createConnectorInitialValues(
   connector: ConfigurableSources
@@ -1180,6 +1186,11 @@ export interface AsanaConfig {
   asana_team_id?: string;
 }
 
+export interface FreshdeskConfig {
+  requested_objects?: string[];
+}
+
+
 export interface MediaWikiConfig extends MediaWikiBaseConfig {
   hostname: string;
 }

web/src/lib/connectors/credentials.ts

@@ -186,6 +186,12 @@ export interface AxeroCredentialJson {
   axero_api_token: string;
 }
 
+export interface FreshdeskCredentialJson {
+  freshdesk_domain: string;
+  freshdesk_password: string;
+  freshdesk_api_key: string;
+}
+
 export interface MediaWikiCredentialJson {}
 export interface WikipediaCredentialJson extends MediaWikiCredentialJson {}
 
@@ -289,6 +295,11 @@ export const credentialTemplates: Record<ValidSources, any> = {
     access_key_id: "",
     secret_access_key: "",
   } as OCICredentialJson,
+  freshdesk: {
+    freshdesk_domain: "",
+    freshdesk_password: "",
+    freshdesk_api_key: "",
+  } as FreshdeskCredentialJson,
   xenforo: null,
   google_sites: null,
   file: null,
@@ -435,6 +446,11 @@ export const credentialDisplayNames: Record<string, string> = {
   // Axero
   base_url: "Axero Base URL",
   axero_api_token: "Axero API Token",
+
+  // Freshdesk
+  freshdesk_domain: "Freshdesk Domain",
+  freshdesk_password: "Freshdesk Password",
+  freshdesk_api_key: "Freshdesk API Key",
 };
 export function getDisplayNameForCredentialKey(key: string): string {
   return credentialDisplayNames[key] || key;

web/src/lib/sources.ts

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

web/src/lib/types.ts

@@ -263,6 +263,7 @@ const validSources = [
   "oci_storage",
   "not_applicable",
   "ingestion_api",
+  "freshdesk",
 ] as const;
 
 export type ValidSources = (typeof validSources)[number];