mirror of
https://github.com/orangecoding/fredy.git
synced 2026-06-16 12:31:07 +00:00
Compare commits
3 Commits
| Author | SHA1 | Date | |
|---|---|---|---|
|
|
db3702ed33 | ||
|
|
e3c62d4696 | ||
|
|
79a8420dfb |
3
index.js
3
index.js
@@ -37,6 +37,9 @@ await runMigrations();
|
||||
// Load provider modules once at startup
|
||||
const providers = await getProviders();
|
||||
|
||||
similarityCache.initSimilarityCache();
|
||||
similarityCache.startSimilarityCacheReloader();
|
||||
|
||||
//assuming interval is always in minutes
|
||||
const INTERVAL = config.interval * 60 * 1000;
|
||||
|
||||
|
||||
@@ -183,8 +183,12 @@ class FredyPipeline {
|
||||
* @returns {Listing[]} Listings considered unique enough to keep.
|
||||
*/
|
||||
_filterBySimilarListings(listings) {
|
||||
const filteredList = listings.filter((listing) => {
|
||||
const similar = this._similarityCache.hasSimilarEntries(listing.title, listing.address);
|
||||
return listings.filter((listing) => {
|
||||
const similar = this._similarityCache.checkAndAddEntry({
|
||||
title: listing.title,
|
||||
address: listing.address,
|
||||
price: listing.price,
|
||||
});
|
||||
if (similar) {
|
||||
logger.debug(
|
||||
`Filtering similar entry for title '${listing.title}' and address '${listing.address}' (Provider: '${this._providerId}')`,
|
||||
@@ -192,8 +196,6 @@ class FredyPipeline {
|
||||
}
|
||||
return !similar;
|
||||
});
|
||||
filteredList.forEach((filter) => this._similarityCache.addCacheEntry(filter.title, filter.address));
|
||||
return filteredList;
|
||||
}
|
||||
|
||||
/**
|
||||
|
||||
@@ -1,3 +1,8 @@
|
||||
### Apprise Adapter
|
||||
|
||||
Refer to the [instructions](https://github.com/caronc/apprise-api#installation) on how to set up an Apprise instance and how to configure your preferred notification service.
|
||||
Use [Apprise](https://github.com/caronc/apprise-api#installation) to forward notifications to many different services.
|
||||
|
||||
Quick start:
|
||||
- Set up an Apprise API instance (see the installation guide linked above).
|
||||
- Configure your preferred notification service(s) within Apprise.
|
||||
- In Fredy, point the Apprise adapter to your Apprise API endpoint.
|
||||
|
||||
@@ -1,4 +1,3 @@
|
||||
### Console Adapter
|
||||
|
||||
The console adapter prints everything found by Fredy into the console (not sending any notifications to you). This can be useful when you want to check if your search
|
||||
criteria meet the expectations.
|
||||
The console adapter prints everything found by Fredy to the console (it does not send notifications). This is useful to verify that your search criteria work as expected before enabling a real notification service.
|
||||
|
||||
@@ -1,4 +1,8 @@
|
||||
### Discord Adapter
|
||||
### Discord Webhook Adapter
|
||||
|
||||
To use the [Discord](https://discord.com/) Adapter, you need to create a webhook on the Discord channel of your choice. You can follow the instructions of _Making A Webhook_ on [this support website](https://support.discord.com/hc/en-us/articles/228383668-Intro-to-Webhooks).
|
||||
Once you have created a webhook, copy and paste the webhook URL.
|
||||
Use a Discord channel webhook to receive notifications.
|
||||
|
||||
Quick start:
|
||||
- Create a webhook in your target Discord channel. See the "Intro to Webhooks" guide on the Discord support site: https://support.discord.com/hc/en-us/articles/228383668-Intro-to-Webhooks
|
||||
- Copy the generated webhook URL.
|
||||
- In Fredy, configure the Discord adapter with this webhook URL.
|
||||
|
||||
@@ -1,8 +1,8 @@
|
||||
### MailJet Adapter
|
||||
### Mailjet Adapter
|
||||
|
||||
To use [MailJet](https://mailjet.com), you need to create an account. You'll need to decide from which email address you want Fredy to send from.
|
||||
To use [Mailjet](https://mailjet.com), create an account and decide which email address Fredy should send from.
|
||||
|
||||
E.g. if you use yourGmailAccount@gmail.com, you have to add this to MailJet and verify it as well.
|
||||
The given public/private api keys are needed in order to use MailJet with Fredy. Fredy will use the same template, it is using for SendGrid.
|
||||
For example, if you use yourGmailAccount@gmail.com, add and verify this address in Mailjet.
|
||||
Provide your public/private API keys in Fredy's configuration. Fredy uses the same email template as for SendGrid.
|
||||
|
||||
If this email should be sent to multiple receiver, use a comma separator (some@email.com, someOther@email.com).
|
||||
To send to multiple recipients, separate email addresses with commas (e.g., some@email.com, someOther@email.com).
|
||||
|
||||
@@ -1,5 +1,8 @@
|
||||
### Mattermost Adapter
|
||||
|
||||
For Mattermost, you need to create a incoming webhook. This is pretty easy. Please visit the steps in the [developer docs](https://docs.mattermost.com/developer/webhooks-incoming.html) and follow the instructions.
|
||||
Receive notifications in Mattermost via an incoming webhook.
|
||||
|
||||
As a result, you get the webhook URL for configuration in fredy. In addition, the target channel must be defined.
|
||||
Quick start:
|
||||
- Create an incoming webhook following the Mattermost developer docs: https://docs.mattermost.com/developer/webhooks-incoming.html
|
||||
- Copy the webhook URL.
|
||||
- In Fredy, configure the Mattermost adapter with this URL and the target channel.
|
||||
|
||||
@@ -1,5 +1,8 @@
|
||||
### ntfy Adapter
|
||||
|
||||
For ntfy, you need to create a topic on your preferred ntfy instance. This is pretty easy. Please visit the steps in the [docs](https://docs.ntfy.sh/publish/) and follow the instructions.
|
||||
Send push notifications using an ntfy topic.
|
||||
|
||||
As a result, you get the URL for configuration in fredy. In addition, the priority must be defined.
|
||||
Quick start:
|
||||
- Create or choose a topic on your preferred ntfy instance (see docs: https://docs.ntfy.sh/publish/).
|
||||
- Copy the publish URL for that topic.
|
||||
- In Fredy, configure the ntfy adapter with the topic URL and set a priority.
|
||||
|
||||
@@ -1,5 +1,8 @@
|
||||
### Pushover Adapter
|
||||
|
||||
Refer to the [instructions](https://support.pushover.net/i7-what-is-pushover-and-how-do-i-use-it) to set up your Pushover application.
|
||||
Use Pushover to receive push notifications on your devices.
|
||||
|
||||
After setting up the application, please enter both your newly created User key and API token.
|
||||
Setup:
|
||||
- Follow Pushover's getting-started guide: https://support.pushover.net/i7-what-is-pushover-and-how-do-i-use-it
|
||||
- Create an application and obtain your User Key and API Token.
|
||||
- In Fredy, configure the Pushover adapter with both values.
|
||||
|
||||
@@ -1,9 +1,12 @@
|
||||
### SendGrid Adapter
|
||||
|
||||
SendGrid is a free email service (free as in "you cannot send more than 100(Sendgrid) and 200(Mailjet) emails a day"), which is more than enough for Fredy.
|
||||
SendGrid is an email delivery service with a generous free tier, which is more than enough for Fredy.
|
||||
|
||||
To use [SendGrid](https://sendgrid.com/), you need to create an account. You'll need to decided from which email address you want Fredy to send from. E.g. if you use yourGmailAccount@gmail.com, you have to add this to sendgrid and verify it as well.
|
||||
Setup:
|
||||
- Create a SendGrid account: https://sendgrid.com/
|
||||
- Decide which email address Fredy should send from (e.g., yourGmailAccount@gmail.com), add it to SendGrid, and complete the verification.
|
||||
- Create an API key and add it to Fredy's configuration.
|
||||
- Create a Dynamic Template in SendGrid. You can copy the template from `/lib/notification/emailTemplate/template.hbs`.
|
||||
|
||||
Lastly you have to create an api-key and feed it into Fredy's config, as well as creating a new dynamic template. For this new template, I recommend copying and pasting the code from the one I have provided under `/lib/notification/emailTemplate/template.hbs`.
|
||||
|
||||
If this email should be sent to multiple receiver use a comma separator (some@email.com, someOther@email.com).
|
||||
Sending to multiple recipients:
|
||||
- Separate email addresses with commas (e.g., some@email.com, someOther@email.com).
|
||||
|
||||
@@ -1,4 +1,5 @@
|
||||
### Slack Adapter
|
||||
IMPORTANT:
|
||||
Don't use this adapter anymore, it is outdated and only here for backwards compatability reasons. Use the new Slack Adapter with webhooks!
|
||||
### Slack Adapter (Legacy)
|
||||
|
||||
*IMPORTANT:*
|
||||
This legacy adapter is outdated and kept only for backward compatibility. Please use the Slack adapter with webhooks instead.
|
||||
|
||||
|
||||
@@ -1,6 +1,10 @@
|
||||
### Slack Adapter
|
||||
### Slack Adapter (Webhooks)
|
||||
|
||||
IMPORTANT:
|
||||
This is the new version of the Slack adapter. I strongly encourage you to use it, the old version is now unmaintained and only kept due to backwards compatability reasons.
|
||||
*IMPORTANT:*
|
||||
This is the recommended Slack adapter. The old Slack adapter is unmaintained and kept only for backward compatibility.
|
||||
|
||||
In order to use [Slack](https://slack.com), you need to create an account. When done, create a new channel and add the Webhook integration to that channel. Copy the webhook url. That's it.
|
||||
Setup:
|
||||
- Create a Slack account and workspace if you don't have one: https://slack.com
|
||||
- Create a channel where you want to receive notifications.
|
||||
- Add the Incoming Webhooks integration to that channel and copy the Webhook URL.
|
||||
- In Fredy, configure the Slack Webhook adapter with this URL.
|
||||
|
||||
@@ -1,9 +1,21 @@
|
||||
### SQLite Adapter
|
||||
|
||||
This adapter stores search results in an SQLite database. By default, the database is located at `db/listings.db`, but you can configure a custom location. This file can be used for further analysis later.
|
||||
This adapter stores search results in an SQLite database. By default, the database is located at `db/listings.db`, but you can configure a custom location. The file can be used for analysis later.
|
||||
|
||||
The database table contains the following columns (all stored as `TEXT` type):
|
||||
The table contains the following columns (all stored as `TEXT`):
|
||||
|
||||
```
|
||||
['serviceName', 'jobKey', 'id', 'size', 'rooms', 'price', 'address', 'title', 'link', 'description', 'image']
|
||||
```json
|
||||
[
|
||||
"serviceName",
|
||||
"jobKey",
|
||||
"id",
|
||||
"size",
|
||||
"rooms",
|
||||
"price",
|
||||
"address",
|
||||
"title",
|
||||
"link",
|
||||
"description",
|
||||
"image"
|
||||
]
|
||||
```
|
||||
|
||||
@@ -117,10 +117,24 @@ export const send = ({ serviceName, newListings = [], notificationConfig, jobKey
|
||||
if (!adapterCfg || !adapterCfg.fields) {
|
||||
throw new Error(`Telegram adapter configuration missing for job '${jobKey || ''}'`);
|
||||
}
|
||||
const { token, chatId } = adapterCfg.fields;
|
||||
const { token, chatId, messageThreadId } = adapterCfg.fields;
|
||||
if (!token || !chatId) {
|
||||
throw new Error("Telegram 'token' and 'chatId' must be provided in notification config");
|
||||
}
|
||||
|
||||
// Optional Telegram topic/thread support (supergroups)
|
||||
let message_thread_id;
|
||||
if (messageThreadId !== undefined && messageThreadId !== null && `${messageThreadId}`.trim() !== '') {
|
||||
const n = Number(messageThreadId);
|
||||
if (Number.isInteger(n) && n > 0) {
|
||||
message_thread_id = n;
|
||||
} else {
|
||||
logger.warn(
|
||||
`Telegram adapter: 'messageThreadId' is invalid ('${messageThreadId}'). It must be a positive integer. Ignoring.`,
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
const job = getJob(jobKey);
|
||||
const jobName = job == null ? jobKey : job.name;
|
||||
|
||||
@@ -147,6 +161,7 @@ export const send = ({ serviceName, newListings = [], notificationConfig, jobKey
|
||||
text: buildText(jobName, serviceName, o),
|
||||
parse_mode: 'HTML',
|
||||
disable_web_page_preview: true,
|
||||
...(message_thread_id ? { message_thread_id } : {}),
|
||||
};
|
||||
|
||||
if (!img) {
|
||||
@@ -160,6 +175,7 @@ export const send = ({ serviceName, newListings = [], notificationConfig, jobKey
|
||||
photo: img,
|
||||
caption: buildCaption(jobName, serviceName, o),
|
||||
parse_mode: 'HTML',
|
||||
...(message_thread_id ? { message_thread_id } : {}),
|
||||
}).catch(async (e) => {
|
||||
logger.error(`Error sending photo to Telegram and use a fallback: ${e.message}`);
|
||||
return await throttledCall('sendMessage', textPayload).catch((e) => {
|
||||
@@ -174,7 +190,7 @@ export const send = ({ serviceName, newListings = [], notificationConfig, jobKey
|
||||
|
||||
/**
|
||||
* Telegram notification adapter configuration schema.
|
||||
* @type {{id:string,name:string,readme:string,description:string,fields:{token:{type:string,label:string,description:string},chatId:{type:string,label:string,description:string}}}}
|
||||
* @type {{id:string,name:string,readme:string,description:string,fields:{token:{type:string,label:string,description:string},chatId:{type:string,label:string,description:string},messageThreadId?:{type:string,label:string,description:string}}}}
|
||||
*/
|
||||
export const config = {
|
||||
id: 'telegram',
|
||||
@@ -192,5 +208,12 @@ export const config = {
|
||||
label: 'Chat Id',
|
||||
description: 'The chat id to send messages to you.',
|
||||
},
|
||||
messageThreadId: {
|
||||
type: 'text',
|
||||
optional: true,
|
||||
label: 'Message Thread Id (optional)',
|
||||
description:
|
||||
'Optional: The topic/thread id within a supergroup to post into (Telegram message_thread_id). Provide a positive integer.',
|
||||
},
|
||||
},
|
||||
};
|
||||
|
||||
@@ -1,12 +1,55 @@
|
||||
### Telegram Adapter
|
||||
|
||||
For Telegram, you need to create a Bot. This is pretty easy. Open [this](https://telegram.me/BotFather) url on your smartphone and follow the instructions.
|
||||
Use this adapter to send notifications to Telegram via a bot. You will need:
|
||||
- A Telegram Bot token (from BotFather)
|
||||
- A chat ID (where messages will be sent)
|
||||
- Optionally: a thread ID if you want to post into a specific forum topic in a group
|
||||
|
||||
A telegram bot is not allowed to send messages directly to a user, you as a user need to first contact the bot to get a chatId.
|
||||
After the user has send a message to your bot the first time, you can gather the chatId like this:
|
||||
#### Create a bot
|
||||
Create a bot with BotFather: open https://telegram.me/BotFather on your phone or in Telegram Desktop and follow the instructions to get your bot token.
|
||||
|
||||
#### Getting the chat ID
|
||||
A Telegram bot cannot message a user first; you must create a conversation (or add the bot to a group/channel) so Telegram assigns a chat the bot can access.
|
||||
|
||||
Steps:
|
||||
1. Start a chat with your bot in Telegram (or add the bot to your group/supergroup/channel) and send any message.
|
||||
2. Fetch recent updates from the Bot API:
|
||||
```
|
||||
curl -X GET "https://api.telegram.org/bot{YOUR_TELEGRAM_TOKEN}/getUpdates"
|
||||
```
|
||||
3. In the JSON response, find the message that you just sent and read `message.chat.id`. That value is your `chatId`.
|
||||
- Private chats: `chat.id` is a positive number
|
||||
- Groups/supergroups: `chat.id` is a negative number
|
||||
|
||||
Keep your bot token secret. If `getUpdates` returns an empty list, send a new message and try again, or make sure your bot’s privacy settings allow it to see group messages when used in groups.
|
||||
|
||||
#### Getting the thread ID (this is optional to be used for forum topics)
|
||||
If you want messages to appear inside a specific forum topic of a supergroup with Topics enabled, you also need a thread ID. In the Telegram Bot API this is called `message_thread_id`.
|
||||
|
||||
When you need it:
|
||||
- Required only for supergroups with Topics enabled when targeting a topic
|
||||
- Not used for private chats, basic groups without Topics, or channels
|
||||
|
||||
Steps to obtain it:
|
||||
1. In your supergroup, enable Topics (Group settings → Manage group → Topics → Enable). Now add a new topic.
|
||||
2. Add your created bot to the topic. (Click on the bot and on "Add to group")
|
||||
3. Open the desired topic (or create a new one) and send any message inside that topic.
|
||||
4. Call `getUpdates` again:
|
||||
```
|
||||
curl -X GET "https://api.telegram.org/bot{YOUR_TELEGRAM_TOKEN}/getUpdates"
|
||||
```
|
||||
4. In the update for the message you sent inside the topic, read `message.message_thread_id`. That number is your `threadId` for this topic.
|
||||
|
||||
Example (truncated):
|
||||
```
|
||||
curl -X GET https://api.telegram.org/bot{YOUR_TELEGRAM_TOKEN}/getUpdates
|
||||
{
|
||||
"message": {
|
||||
"chat": { "id": -1001234567890, "type": "supergroup" },
|
||||
"message_thread_id": 42,
|
||||
"text": "hello from the topic"
|
||||
}
|
||||
}
|
||||
```
|
||||
Use `chat.id` as `chatId` and `message_thread_id` as `threadId` in your configuration.
|
||||
|
||||
A more detailed list of instructions can be found here [https://core.telegram.org/bots#botfather](https://core.telegram.org/bots#botfather)
|
||||
More details about bots and BotFather: https://core.telegram.org/bots#botfather
|
||||
|
||||
@@ -1,6 +1,4 @@
|
||||
import markdown$0 from 'markdown';
|
||||
import fs from 'fs';
|
||||
const markdown = markdown$0.markdown;
|
||||
export function markdown2Html(filePath) {
|
||||
return markdown.toHTML(fs.readFileSync(filePath, 'utf8'));
|
||||
return fs.readFileSync(filePath, 'utf8');
|
||||
}
|
||||
|
||||
@@ -1,116 +1,94 @@
|
||||
import crypto from 'crypto';
|
||||
|
||||
const retention = 60 * 60 * 1000;
|
||||
/**
|
||||
* Internal cache storage.
|
||||
* Maps a SHA-256 hash (string) to its expiry timestamp (number in ms).
|
||||
* @type {Map<string, number>}
|
||||
*/
|
||||
const entries = new Map();
|
||||
|
||||
/**
|
||||
* Reference to the currently scheduled cleanup timer.
|
||||
* @type {NodeJS.Timeout | null}
|
||||
*/
|
||||
let timer = null;
|
||||
|
||||
/**
|
||||
* Generate a SHA-256 hash from a list of input strings.
|
||||
* Null or undefined values are ignored.
|
||||
* Similarity cache
|
||||
*
|
||||
* @param {...(string|null|undefined)} strings - Input values to hash
|
||||
* Maintains an in-memory Set of content hashes to detect whether a listing
|
||||
* (identified by a tuple of title, price and address) has been seen before.
|
||||
*
|
||||
* Design notes:
|
||||
* - The cache is refreshed periodically from persistent storage. To avoid
|
||||
* modification-during-iteration issues, the refresh builds a new Set and
|
||||
* atomically swaps the reference instead of mutating in place.
|
||||
* - Hashing ignores null/undefined values but preserves falsy-yet-valid values
|
||||
* like 0. Non-string values are coerced to strings before hashing.
|
||||
*
|
||||
* This module has no persistence of its own; it relies on
|
||||
* getAllEntriesFromListings() for data hydration.
|
||||
* @module similarityCache
|
||||
*/
|
||||
import crypto from 'crypto';
|
||||
import { getAllEntriesFromListings } from '../storage/listingsStorage.js';
|
||||
|
||||
/** @type {number} Refresh interval in milliseconds (defaults to one hour). */
|
||||
const reloadCycle = 60 * 60 * 1000; // every hour, refresh
|
||||
|
||||
/**
|
||||
* Internal cache of content hashes for known listings.
|
||||
*
|
||||
* Each entry is an SHA-256 hex digest produced by toHash(title, price, address).
|
||||
* @type {Set<string>}
|
||||
*/
|
||||
let cache = new Set();
|
||||
|
||||
export const startSimilarityCacheReloader = () => {
|
||||
// Periodically refresh the cache from storage
|
||||
setInterval(() => {
|
||||
initSimilarityCache();
|
||||
}, reloadCycle);
|
||||
};
|
||||
|
||||
/**
|
||||
* Initialize or refresh the similarity cache from persistent storage.
|
||||
*
|
||||
* Reads all stored listings via getAllEntriesFromListings(), computes a hash for
|
||||
* each, and swaps the in-memory Set atomically to avoid in-place mutations that
|
||||
* could interfere with concurrent iteration.
|
||||
*
|
||||
* This function is idempotent and safe to call at any time.
|
||||
* @returns {void}
|
||||
*/
|
||||
export const initSimilarityCache = () => {
|
||||
const allEntries = getAllEntriesFromListings();
|
||||
const newCache = new Set();
|
||||
for (const entry of allEntries) {
|
||||
newCache.add(toHash(entry?.title, entry?.price, entry?.address));
|
||||
}
|
||||
// Atomic swap to avoid mutating the cache while it may be iterated elsewhere
|
||||
cache = newCache;
|
||||
};
|
||||
|
||||
/**
|
||||
* Check if a listing is already known and add it to the cache if not.
|
||||
*
|
||||
* The listing is identified by the combination of its title, price and
|
||||
* address. Null/undefined fields are ignored during hashing. Falsy-but-valid
|
||||
* values (e.g., price 0) are preserved.
|
||||
*
|
||||
* @param {Object} params - Listing fields
|
||||
* @param {string|undefined|null} params.title - The listing title
|
||||
* @param {string|undefined|null} params.address - The listing address
|
||||
* @param {number|string|undefined|null} params.price - The listing price
|
||||
* @returns {boolean} true if the entry already existed in the cache (duplicate), otherwise false
|
||||
*/
|
||||
export const checkAndAddEntry = ({ title, address, price }) => {
|
||||
const hash = toHash(title, price, address);
|
||||
if (cache.has(hash)) {
|
||||
return true;
|
||||
}
|
||||
cache.add(hash);
|
||||
return false;
|
||||
};
|
||||
|
||||
/**
|
||||
* Generate an SHA-256 hash from a list of input values.
|
||||
* Null or undefined values are ignored. Falsy but valid values like 0 are preserved.
|
||||
* Non-string values are coerced to strings prior to hashing.
|
||||
*
|
||||
* @param {...(string|number|null|undefined)} strings - Input values to hash
|
||||
* @returns {string} Hexadecimal hash
|
||||
*/
|
||||
function toHash(...strings) {
|
||||
return crypto.createHash('sha256').update(strings.filter(Boolean).join('|')).digest('hex');
|
||||
}
|
||||
|
||||
/**
|
||||
* Cleanup expired cache entries and schedule the next cleanup run.
|
||||
* This function is invoked automatically by scheduled timers.
|
||||
*
|
||||
* @private
|
||||
*/
|
||||
function runCleanup() {
|
||||
const now = Date.now();
|
||||
for (const [hash, expiry] of entries) {
|
||||
if (expiry <= now) entries.delete(hash);
|
||||
}
|
||||
scheduleNext();
|
||||
}
|
||||
|
||||
/**
|
||||
* Find the soonest expiry timestamp among all cache entries
|
||||
* and schedule a one-shot timer that will trigger at that time.
|
||||
* Cancels any existing timer before scheduling a new one.
|
||||
*
|
||||
* @private
|
||||
*/
|
||||
function scheduleNext() {
|
||||
if (timer) {
|
||||
clearTimeout(timer);
|
||||
timer = null;
|
||||
}
|
||||
let next = Infinity;
|
||||
const now = Date.now();
|
||||
for (const expiry of entries.values()) {
|
||||
if (expiry > now && expiry < next) next = expiry;
|
||||
}
|
||||
if (next !== Infinity) {
|
||||
timer = setTimeout(runCleanup, Math.max(0, next - now));
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Add or refresh a cache entry for the given title and address.
|
||||
* The entry will automatically expire after the configured retention window.
|
||||
*
|
||||
* @param {string} title - The title used to build the cache key
|
||||
* @param {string} address - The address used to build the cache key
|
||||
*/
|
||||
export function addCacheEntry(title, address) {
|
||||
const hash = toHash(title, address);
|
||||
const expiry = Date.now() + retention;
|
||||
entries.set(hash, expiry);
|
||||
scheduleNext();
|
||||
}
|
||||
|
||||
/**
|
||||
* Check if a cache entry with the same title and address exists
|
||||
* and is still valid (not expired).
|
||||
*
|
||||
* @param {string} title - The title used to build the cache key
|
||||
* @param {string} address - The address used to build the cache key
|
||||
* @returns {boolean} True if a valid cache entry exists, false otherwise
|
||||
*/
|
||||
export function hasSimilarEntries(title, address) {
|
||||
const hash = toHash(title, address);
|
||||
const expiry = entries.get(hash);
|
||||
if (expiry == null) return false;
|
||||
if (expiry <= Date.now()) {
|
||||
entries.delete(hash);
|
||||
scheduleNext();
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* Stop any scheduled cleanup timers and prevent further automatic cleanup.
|
||||
* Entries that are already in the cache will remain until removed manually
|
||||
* or until cleanup is started again by adding new entries.
|
||||
*/
|
||||
export function stopCacheCleanup() {
|
||||
if (timer) clearTimeout(timer);
|
||||
timer = null;
|
||||
}
|
||||
|
||||
/**
|
||||
* this is only for test purposes
|
||||
*/
|
||||
export function invalidateAllForTest() {
|
||||
for (const key of entries.keys()) {
|
||||
entries.set(key, 0);
|
||||
}
|
||||
runCleanup();
|
||||
const normalized = strings
|
||||
.filter((v) => v !== null && v !== undefined)
|
||||
.map((v) => (typeof v === 'string' ? v : String(v)));
|
||||
return crypto.createHash('sha256').update(normalized.join('|')).digest('hex');
|
||||
}
|
||||
|
||||
@@ -310,8 +310,8 @@ export const deleteListingsByJobId = (jobId) => {
|
||||
if (!jobId) return;
|
||||
return SqliteConnection.execute(
|
||||
`DELETE
|
||||
FROM listings
|
||||
WHERE job_id = @jobId`,
|
||||
FROM listings
|
||||
WHERE job_id = @jobId`,
|
||||
{ jobId },
|
||||
);
|
||||
};
|
||||
@@ -332,3 +332,13 @@ export const deleteListingsById = (ids) => {
|
||||
ids,
|
||||
);
|
||||
};
|
||||
|
||||
/**
|
||||
* Return all listings with only the fields: title, address, and price.
|
||||
* This is the single helper requested for simple consumers.
|
||||
*
|
||||
* @returns {{title: string|null, address: string|null, price: number|null}[]}
|
||||
*/
|
||||
export const getAllEntriesFromListings = () => {
|
||||
return SqliteConnection.query(`SELECT title, address, price FROM listings`);
|
||||
};
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "fredy",
|
||||
"version": "14.2.2",
|
||||
"version": "14.3.1",
|
||||
"description": "[F]ind [R]eal [E]states [d]amn eas[y].",
|
||||
"scripts": {
|
||||
"prepare": "husky",
|
||||
@@ -69,7 +69,6 @@
|
||||
"cookie-session": "2.1.1",
|
||||
"handlebars": "4.7.8",
|
||||
"lodash": "4.17.21",
|
||||
"markdown": "^0.5.0",
|
||||
"nanoid": "5.1.6",
|
||||
"node-cron": "^4.2.1",
|
||||
"node-fetch": "3.3.2",
|
||||
|
||||
@@ -1,53 +0,0 @@
|
||||
import { expect } from 'chai';
|
||||
import * as similarityCache from '../../lib/services/similarity-check/similarityCache.js';
|
||||
import { mockFredy } from '../utils.js';
|
||||
|
||||
describe('FredyPipeline', () => {
|
||||
afterEach(() => {
|
||||
similarityCache.invalidateAllForTest();
|
||||
});
|
||||
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
describe('_filterBySimilarListings', () => {
|
||||
let fredyRuntime;
|
||||
|
||||
beforeEach(async () => {
|
||||
const FredyRuntime = await mockFredy();
|
||||
fredyRuntime = new FredyRuntime({}, null, 'dummy-provider', 'dummy-job', similarityCache);
|
||||
});
|
||||
|
||||
it('should filter out listings with similar title and address already in cache', () => {
|
||||
similarityCache.addCacheEntry('Penthouse', 'Mustermann Straße 1');
|
||||
|
||||
const listings = [
|
||||
{ id: '1', title: 'Penthouse', address: 'Mustermann Straße 1' },
|
||||
{ id: '2', title: 'Nice apartment', address: 'Mustermann Straße 15' },
|
||||
];
|
||||
|
||||
const result = fredyRuntime._filterBySimilarListings(listings);
|
||||
|
||||
expect(result).to.have.length(1);
|
||||
expect(result[0].id).to.equal('2');
|
||||
expect(result[0].title).to.equal('Nice apartment');
|
||||
|
||||
expect(similarityCache.hasSimilarEntries('Nice apartment', 'Mustermann Straße 15')).to.be.true;
|
||||
});
|
||||
|
||||
it('should handle listings with null or undefined address', () => {
|
||||
const listings = [
|
||||
{ id: '1', title: 'Penthouse', address: null },
|
||||
{ id: '2', title: 'Nice apartment', address: undefined },
|
||||
];
|
||||
|
||||
const result = fredyRuntime._filterBySimilarListings(listings);
|
||||
|
||||
expect(result).to.have.length(2);
|
||||
|
||||
expect(similarityCache.hasSimilarEntries('Penthouse', null)).to.be.true;
|
||||
expect(similarityCache.hasSimilarEntries('Nice apartment', undefined)).to.be.true;
|
||||
});
|
||||
});
|
||||
});
|
||||
@@ -5,9 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/einsAImmobilien.js';
|
||||
|
||||
describe('#einsAImmobilien testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
provider.init(providerConfig.einsAImmobilien, [], []);
|
||||
it('should test einsAImmobilien provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
|
||||
@@ -5,9 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/immobilienDe.js';
|
||||
|
||||
describe('#immobilien.de testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
provider.init(providerConfig.immobilienDe, [], []);
|
||||
it('should test immobilien.de provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
|
||||
@@ -5,10 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/immonet.js';
|
||||
|
||||
describe('#immonet testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
it('should test immonet provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
provider.init(providerConfig.immonet, [], []);
|
||||
|
||||
@@ -5,10 +5,6 @@ import { get } from '../mocks/mockNotification.js';
|
||||
import * as provider from '../../lib/provider/immoscout.js';
|
||||
|
||||
describe('#immoscout provider testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
provider.init(providerConfig.immoscout, [], []);
|
||||
it('should test immoscout provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
|
||||
@@ -5,9 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/immoswp.js';
|
||||
|
||||
describe('#immoswp testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
provider.init(providerConfig.immoswp, [], []);
|
||||
it('should test immoswp provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
|
||||
@@ -5,10 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/immowelt.js';
|
||||
|
||||
describe('#immowelt testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
it('should test immowelt provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
provider.init(providerConfig.immowelt, [], []);
|
||||
|
||||
@@ -5,9 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/kleinanzeigen.js';
|
||||
|
||||
describe('#kleinanzeigen testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
it('should test kleinanzeigen provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
provider.init(providerConfig.kleinanzeigen, [], []);
|
||||
|
||||
@@ -5,10 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/mcMakler.js';
|
||||
|
||||
describe('#mcMakler testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
it('should test mcMakler provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
provider.init(providerConfig.mcMakler, []);
|
||||
|
||||
@@ -5,9 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/neubauKompass.js';
|
||||
|
||||
describe('#neubauKompass testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
provider.init(providerConfig.neubauKompass, [], []);
|
||||
it('should test neubauKompass provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
|
||||
@@ -5,10 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/regionalimmobilien24.js';
|
||||
|
||||
describe('#regionalimmobilien24 testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
it('should test regionalimmobilien24 provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
provider.init(providerConfig.regionalimmobilien24, []);
|
||||
|
||||
@@ -5,10 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/sparkasse.js';
|
||||
|
||||
describe('#sparkasse testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
|
||||
it('should test sparkasse provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
provider.init(providerConfig.sparkasse, []);
|
||||
|
||||
@@ -5,9 +5,6 @@ import { expect } from 'chai';
|
||||
import * as provider from '../../lib/provider/wgGesucht.js';
|
||||
|
||||
describe('#wgGesucht testsuite()', () => {
|
||||
after(() => {
|
||||
similarityCache.stopCacheCleanup();
|
||||
});
|
||||
provider.init(providerConfig.wgGesucht, [], []);
|
||||
it('should test wgGesucht provider', async () => {
|
||||
const Fredy = await mockFredy();
|
||||
|
||||
@@ -1,30 +0,0 @@
|
||||
import { expect } from 'chai';
|
||||
import * as similarityCache from '../../lib/services/similarity-check/similarityCache.js';
|
||||
|
||||
describe('similarityCheck', () => {
|
||||
it('should return true on duplicate', () => {
|
||||
similarityCache.addCacheEntry('Hello World', 'Test');
|
||||
expect(similarityCache.hasSimilarEntries('Hello World', 'Test')).to.be.true;
|
||||
});
|
||||
|
||||
it('should return true even if one value is null', () => {
|
||||
similarityCache.addCacheEntry('Hello World', null);
|
||||
expect(similarityCache.hasSimilarEntries('Hello World', null)).to.be.true;
|
||||
});
|
||||
|
||||
it('should return true even if one value is an obj', () => {
|
||||
similarityCache.addCacheEntry('Hello World', [{ TR: 'OLOLO' }]);
|
||||
expect(similarityCache.hasSimilarEntries('Hello World', [{ TR: 'OLOLO' }])).to.be.true;
|
||||
});
|
||||
|
||||
it('should return false when no duplicate', () => {
|
||||
similarityCache.addCacheEntry('Hello World__', 'Test');
|
||||
expect(similarityCache.hasSimilarEntries('Hello World___', 'Test')).to.be.false;
|
||||
});
|
||||
|
||||
it('should return false when no duplicate', () => {
|
||||
expect(similarityCache.hasSimilarEntries('Hello World', 'Test')).to.be.true;
|
||||
similarityCache.invalidateAllForTest();
|
||||
expect(similarityCache.hasSimilarEntries('Hello World', 'Test')).to.be.false;
|
||||
});
|
||||
});
|
||||
62
test/similarity/similarityCache.test.js
Normal file
62
test/similarity/similarityCache.test.js
Normal file
@@ -0,0 +1,62 @@
|
||||
import { expect } from 'chai';
|
||||
import esmock from 'esmock';
|
||||
|
||||
// Helper to create module under test with mocks
|
||||
async function loadModuleWith({ entries = [] } = {}) {
|
||||
const mod = await esmock('../../lib/services/similarity-check/similarityCache.js', {
|
||||
// Mock the storage to return our controlled entries
|
||||
'../../lib/services/storage/listingsStorage.js': {
|
||||
getAllEntriesFromListings: () => entries,
|
||||
},
|
||||
});
|
||||
return mod;
|
||||
}
|
||||
|
||||
describe('similarityCache', () => {
|
||||
it('initSimilarityCache builds cache from storage and enables duplicate detection', async () => {
|
||||
const entries = [
|
||||
{ title: 'A', price: 1000, address: 'Main 1' },
|
||||
{ title: 'B', price: 0, address: 'Zero St' },
|
||||
];
|
||||
|
||||
const { initSimilarityCache, checkAndAddEntry } = await loadModuleWith({ entries });
|
||||
|
||||
// Initially, duplicates should not be detected for new data
|
||||
expect(checkAndAddEntry({ title: 'X', price: 200, address: 'Y' })).to.equal(false);
|
||||
|
||||
// Now initialize from storage
|
||||
initSimilarityCache();
|
||||
|
||||
// Exact duplicates should be detected
|
||||
expect(checkAndAddEntry({ title: 'A', price: 1000, address: 'Main 1' })).to.equal(true);
|
||||
// Ensure falsy-but-valid price 0 is preserved by hashing and detected as duplicate
|
||||
expect(checkAndAddEntry({ title: 'B', price: 0, address: 'Zero St' })).to.equal(true);
|
||||
});
|
||||
|
||||
it('checkAndAddEntry returns false for new entry then true for duplicate on second call', async () => {
|
||||
const { checkAndAddEntry } = await loadModuleWith();
|
||||
|
||||
const first = checkAndAddEntry({ title: 'C', price: 300, address: 'Road 3' });
|
||||
const second = checkAndAddEntry({ title: 'C', price: 300, address: 'Road 3' });
|
||||
|
||||
expect(first).to.equal(false);
|
||||
expect(second).to.equal(true);
|
||||
});
|
||||
|
||||
it('hashing ignores null/undefined but preserves 0 via behavior', async () => {
|
||||
const { checkAndAddEntry } = await loadModuleWith();
|
||||
|
||||
// Add baseline (null address ignored)
|
||||
const add1 = checkAndAddEntry({ title: 'T', price: 1, address: null });
|
||||
expect(add1).to.equal(false);
|
||||
// Duplicate with undefined address should match
|
||||
const dup = checkAndAddEntry({ title: 'T', price: 1, address: undefined });
|
||||
expect(dup).to.equal(true);
|
||||
|
||||
// Now test that price 0 is preserved (not filtered out)
|
||||
const addZero = checkAndAddEntry({ title: 'Z', price: 0, address: 'Zero' });
|
||||
expect(addZero).to.equal(false);
|
||||
const dupZero = checkAndAddEntry({ title: 'Z', price: 0, address: 'Zero' });
|
||||
expect(dupZero).to.equal(true);
|
||||
});
|
||||
});
|
||||
@@ -21,7 +21,7 @@ const sortAdapter = (a, b) => {
|
||||
const validate = (selectedAdapter) => {
|
||||
const results = [];
|
||||
for (let uiElement of Object.values(selectedAdapter.fields || [])) {
|
||||
if (uiElement.value == null) {
|
||||
if (uiElement.value == null && !uiElement.optional) {
|
||||
results.push('All fields are mandatory and must be set.');
|
||||
continue;
|
||||
}
|
||||
@@ -36,7 +36,7 @@ const validate = (selectedAdapter) => {
|
||||
results.push('A boolean field cannot be of a different type.');
|
||||
continue;
|
||||
}
|
||||
if (typeof uiElement.value === 'string' && uiElement.value.length === 0) {
|
||||
if (typeof uiElement.value === 'string' && uiElement.value.length === 0 && !uiElement.optional) {
|
||||
results.push('All fields are mandatory and must be set.');
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
import React from 'react';
|
||||
import { Banner } from '@douyinfe/semi-ui';
|
||||
import { Banner, MarkdownRender } from '@douyinfe/semi-ui';
|
||||
|
||||
export default function Help({ readme }) {
|
||||
return (
|
||||
@@ -8,7 +8,7 @@ export default function Help({ readme }) {
|
||||
type="info"
|
||||
closeIcon={null}
|
||||
title={<div style={{ fontWeight: 600, fontSize: '14px', lineHeight: '20px' }}>Information</div>}
|
||||
description={<p dangerouslySetInnerHTML={{ __html: readme }} />}
|
||||
description={<MarkdownRender raw={readme} />}
|
||||
/>
|
||||
);
|
||||
}
|
||||
|
||||
19
yarn.lock
19
yarn.lock
@@ -1909,11 +1909,6 @@
|
||||
"@types/babel__core" "^7.20.5"
|
||||
react-refresh "^0.17.0"
|
||||
|
||||
abbrev@1:
|
||||
version "1.1.1"
|
||||
resolved "https://registry.yarnpkg.com/abbrev/-/abbrev-1.1.1.tgz#f8f2c887ad10bf67f634f005b6987fed3179aac8"
|
||||
integrity sha512-nne9/IiQ/hzIhY6pdDnbBtz7DjPTKrY00P/zvPSm5pOFkl6xuGrGnXn/VtTNNfNtAfZ9/1RtehkszU9qcTii0Q==
|
||||
|
||||
abs-svg-path@~0.1.1:
|
||||
version "0.1.1"
|
||||
resolved "https://registry.yarnpkg.com/abs-svg-path/-/abs-svg-path-0.1.1.tgz#df601c8e8d2ba10d4a76d625e236a9a39c2723bf"
|
||||
@@ -4689,13 +4684,6 @@ markdown-table@^3.0.0:
|
||||
resolved "https://registry.yarnpkg.com/markdown-table/-/markdown-table-3.0.4.tgz#fe44d6d410ff9d6f2ea1797a3f60aa4d2b631c2a"
|
||||
integrity sha512-wiYz4+JrLyb/DqW2hkFJxP7Vd7JuTDm77fvbM8VfEQdmSMqcImWeeRbHwZjBjIFki/VaMK2BhFi7oUUZeM5bqw==
|
||||
|
||||
markdown@^0.5.0:
|
||||
version "0.5.0"
|
||||
resolved "https://registry.yarnpkg.com/markdown/-/markdown-0.5.0.tgz#28205b565a8ae7592de207463d6637dc182722b2"
|
||||
integrity sha512-ctGPIcuqsYoJ493sCtFK7H4UEgMWAUdXeBhPbdsg1W0LsV9yJELAHRsMmWfTgao6nH0/x5gf9FmsbxiXnrgaIQ==
|
||||
dependencies:
|
||||
nopt "~2.1.1"
|
||||
|
||||
math-intrinsics@^1.1.0:
|
||||
version "1.1.0"
|
||||
resolved "https://registry.yarnpkg.com/math-intrinsics/-/math-intrinsics-1.1.0.tgz#a0dd74be81e2aa5c2f27e65ce283605ee4e2b7f9"
|
||||
@@ -5508,13 +5496,6 @@ nodemon@^3.1.10:
|
||||
touch "^3.1.0"
|
||||
undefsafe "^2.0.5"
|
||||
|
||||
nopt@~2.1.1:
|
||||
version "2.1.2"
|
||||
resolved "https://registry.yarnpkg.com/nopt/-/nopt-2.1.2.tgz#6cccd977b80132a07731d6e8ce58c2c8303cf9af"
|
||||
integrity sha512-x8vXm7BZ2jE1Txrxh/hO74HTuYZQEbo8edoRcANgdZ4+PCV+pbjd/xdummkmjjC7LU5EjPzlu8zEq/oxWylnKA==
|
||||
dependencies:
|
||||
abbrev "1"
|
||||
|
||||
normalize-path@^3.0.0, normalize-path@~3.0.0:
|
||||
version "3.0.0"
|
||||
resolved "https://registry.yarnpkg.com/normalize-path/-/normalize-path-3.0.0.tgz#0dcd69ff23a1c9b11fd0978316644a0388216a65"
|
||||
|
||||
Reference in New Issue
Block a user