Hello everyone, I am LINE Tech Evangelist - Evan Lin. LINE is committed to holding internal technical exchanges, training, and external community gatherings, campus speeches, developer recruitment days, developer conferences, and various other activities. We hope to create more opportunities for technical sharing and cross-national exchanges, and at the same time, continuously recruit outstanding talents to join the LINE Taiwan development engineering team!
Tonight, I participated in the innovative Test Corner meetup and explored the mysteries of testing with the attendees. The various ingenious ideas of this gathering were refreshing, and the atmosphere full of community vitality made me feel passionate as well. Let's experience the charm of this gathering together!
KKTIX event webpage: Event URL
Tonight's Test Corner meetup was specially opened by the sweet and lovely host – Sherry, with her humorous (like a flight attendant showing the surroundings) opening. She also introduced the LINE TECH FRESH campus new star talent program. What is LINE TECH FRESH? The LINE Taiwan engineering team recruits students/graduate students from information technology-related departments or those who have some knowledge in this field through the LINE TECH FRESH – Technical New Star Talent Program every year to join the LINE team for long-term internships (one-year term), allowing students to observe and learn in an international technology company.
For more information, you can refer to this article: LINE TECH FRESH – Technical New Star Talent Program, Internship Experience Revealed
LINE SPOT Mask Map - Vince Chen
LINE SPOT is a platform that connects physical businesses with life services, allowing users to instantly check for discounts and services near their location. LINE SPOT puts these services together, deepening the connection between the online/offline ecosystem. And such convenient services naturally face many challenges during the development process, including integrating many different services, frequent service migration, and supporting high traffic for the system.
As an Automation Engineer, the main job content in this team is divided into (1) API verification, confirming the correctness of the data. (2) System performance tuning and optimization.
The traffic when the mask map was launched exceeded 30 times more than originally expected. So how does LINE SPOT do similar related testing?
First, the LINE SPOT team uses the k6.io tool, along with the Jenkins pipeline process. Every time there is a modification online, it will be deployed to the test machine, and k6.io will be used for load testing and performance testing. Each result will be preset to be thrown into InfluxDB for storage, and Grafana can be used to visualize the entire test performance report. This allows testers and developers to see the entire result at a glance.
Of course, for such a large amount of traffic, the entire service has also done a lot of performance tuning.
- For the API Gateway, there is a Query Efficiency check to check that each call is correct and ensure query performance.
- For ElasticSearch, there is also Connection Pool Tuning, which provides more protection for the use of query resources (no excessive query connections will be opened).
- There is also cache and call monitor for the API Gateway to monitor whether performance is affected.
Lesson learned from Mask Map
The LINE SPOT team learned the following related matters after experiencing the development and operation of the "Mask Map" related functions:
- Monitoring and performance management are very important for Micro-services
- Always pay attention to performance optimization and changes after adjustment
- Through some customized small tools, the team can be more convenient in management and operation.
LINE Message Verification - Hank Chen and Johnson Wu
This section is shared by Hank Chen and Johnson Wu, divided into two parts. The first part is Hank's presentation on the LINE message verification test process and how to do Test content and Test Case well. The second part is Johnson Wu's presentation on the message verification data processing process and data processing platform.
Part 1: LINE Message Verification Test Process
The "LINE Message Verification" platform service was developed. The "LINE Message Verification" platform officially launched last July, not only has an official website, but also connects to the LINE official account. Users only need to "forward" the messages received in the chat room to the "LINE Message Verification" official account. If there are already verification reports in the database, the verification assistant will automatically determine its authenticity, and the system will provide the verification results immediately; if the message has not been verified, it will be reported to a professional verification unit, and after clarification, the correct information will be returned to the user as soon as possible, providing the most immediate message identification service, helping users identify the authenticity of suspicious messages, and reducing the chance of false messages being spread again.
The Agile test process for message verification is as follows:
- Planning Refinement:
- Acceptance Criteria:
- Functional Test:
- Regression Test:
- Exploratory Test:
- Release Retrospective:
In the entire Agile process, there is a 15-minute standup meeting every day to ensure that the members of the project can start working under the same consensus.
And for the establishment of content detection (Test Case establishment), the goal is to have a clear process and a clear description. Colleagues with the same skills can quickly continue testing. The overall testing platform is divided into manual testing and automated testing, and the TestRail platform is used for automated testing.
CICD Release Process:
Using Drone and git, as soon as there is related code submission, it will be deployed to Kubernetes through Argo.
New features of the message verification assistant
- Can query problematic official accounts
- Can quickly check problematic URLs
Interested readers are welcome to join the LINE Message Verification Assistant official account
Part 2: How to do data processing and analysis in "LINE Message Verification"
First, how can machine learning help with message verification? The main methods are as follows:
- Use Similarity Search to find similar false messages and directly reply to users.
- And classify similar false messages through Classification and Near-Duplication.
Among them, Near-Duplication is used to calculate whether the received message is similar to the known false message through Near-Duplication. If it is confirmed to be similar, or partially similar. It means that the probability of being a false message will also increase.
And the complete process will first use Ner-Duplication for comparison to find possible approximate false messages for users to confirm. After confirmation, the user's query message will also be classified in the category of the false message.
Among them, the entire system also includes mlFlow, which controls the machine computing model process, and airFlow, which is used as the data processing process control. The trained strings and labels are scheduled and trained through airFlow, and then the GPU is used to accelerate the training results. Through mlFlow, the calculated model can be deployed to the Predictor Server for query.
Finally, the speaker also shared the relevant project results data:
- 46% of users have identified fake news
- Among them, 33% of users are quite active in checking
- And 25% will share the results with friends
This is really the only way to prevent false messages as soon as possible, and not let you and me become accomplices of false messages.
Event Summary
After a year, we are able to invite Test Corner to the LINE office again. I am very happy to see that many student friends have a high interest in LINE's product testing. LINE, as a part of high self-discipline and social responsibility, in addition to allowing more users to obtain the most immediate and local information through LINE SPOT, also hopes to allow every user to identify the authenticity of messages as soon as possible through "Message Verification" and avoid becoming an accomplice.
Join the "LINE Developer Official Community" official account immediately, and you can receive the first-hand Meetup activities, or push notifications of the latest news related to the developer program. ▼
"LINE Developer Official Community" official account ID: @line_tw_dev
About the "LINE Developer Community Program"
LINE launched the "LINE Developer Community Program" in Taiwan at the beginning of this year, and will invest long-term manpower and resources in Taiwan to hold internal and external developer community gatherings, recruitment days, developer conferences, etc., and has held more than 30 events. Readers are welcome to continue to check back for the latest updates. For details, please see:
- 2019 LINE Developer Community Program Activity Schedule
- LINE Taiwan Developer Relations 2019 Review and 2019 Developer Community Program Report
- 2020 LINE Developer Community Program Activity Schedule
Recruitment Information
《LINE is strongly recruiting!》Join us to Close the Distance and connect the smart new world » Detailed job information
Top comments (0)