1 views collected around this technical thread.
This tutorial explains how to use Python 3 with Requests, lxml, and Baidu's AipFace SDK to crawl images from Zhihu's "美女" topic, filter them by face detection, gender, authenticity, and beauty score, and store the qualified pictures locally.
This article presents a method for massive multi-label text classification on Zhihu content by combining a semantic retrieval model with a proprietary large AI model, detailing the challenges of large label spaces, model architecture, loss optimization, and experimental results showing significant accuracy gains.
This article presents a comprehensive overview of Zhihu's Bridge platform, detailing its content‑operation architecture—including content pool, management, analysis, monitoring, and intervention modules—explaining the underlying streaming and batch technologies such as Flink, Doris, and Elasticsearch, and outlining future automation and AI‑driven workflow directions.
This article explains how to collect images from Zhihu's "beauty" topic using Python's Requests and lxml libraries, filter them with Baidu AI's AipFace face detection service, and store the qualified pictures locally, detailing the required environment, logic, and preparation steps.
This article demonstrates how to automate Zhihu login on Windows by analyzing network traffic with Fiddler, extracting required parameters, and implementing a Python script that builds HTTP requests using urllib2, handles cookies, captcha retrieval, and logs the results, complete with sample code and execution screenshots.
This tutorial demonstrates how to build a Scrapy spider that crawls Zhihu user follow and follower data via Zhihu’s public APIs, handles request headers, parses JSON responses, paginates results, and stores the extracted information into MongoDB using a custom item pipeline.
This article presents a comprehensive overview of Zhihu's recommendation page ranking system, detailing the request flow, ranking evolution from time‑based to deep‑learning models, feature engineering strategies, model architectures such as DNN, DeepFM, DIN, multi‑task learning, and lessons learned for production deployment.
This article details Zhihu's read‑service architecture, covering its business requirements, high‑availability and high‑performance design goals, key components such as Proxy, Cache and Storage, extensive performance metrics, the migration from MySQL to TiDB, and the benefits brought by TiDB 3.0 features.
This tutorial demonstrates how to use Python's requests library to simulate Zhihu login by handling dynamic _xsrf tokens, optional captcha verification, saving cookies, and then crawling the main page to extract questions and answer abstracts.
In a playful data‑driven experiment, the author scraped 27,664 Zhihu answers to “What are your dating criteria?”, filtered out short, outdated, high‑profile or already‑matched posts, applied follower‑and engagement‑thresholds to narrow the pool to 480 candidates, then ranked the top 30 by a like‑to‑comment ratio, sharing the code and dataset for reproducibility.
This article walks through a complete data‑analysis workflow—scraping Zhihu dating‑preference answers, cleaning and filtering the data, deriving gender and activity metrics, designing a four‑step screening process, and finally ranking candidates with a custom like‑to‑comment index—to help a single programmer create a concise, high‑quality list of potential partners.
This article shows how to use Python to scrape Zhihu answers, store them in MongoDB, filter for short high‑upvote replies, and then presents a collection of programmer‑centric jokes that illustrate the kind of "god replies" the crawler can retrieve.