用于计算给定数据集的平均数、标准差以及两组数据间的P值。当用户要求“无需过程”或“只需结果”时，仅输出最终数值。

用于将Pandas DataFrame同步到MySQL数据库的技能，包含针对JSON字段、数值类型（int/float）、字符串顺序（如effect_desc）的归一化比较逻辑，以及基于type字段的条件排除规则。

用于在Pandas DataFrame执行合并操作后，清理由merge产生的后缀列（_x, _y）和指示列，使其符合数据库表结构以便进行to_sql插入操作。

从Excel读取中文文本，支持基于句法树结构（如深度、词性）或语义向量（如TF-IDF、SentenceTransformer）的特征提取，执行聚类分析（如K-Means、DBSCAN），并将结果导出回Excel。

使用Python和pandas读取Excel文件，对比“依存关系”和“识别的依存关系”列，在剔除特定符号（如?）后计算正确率和错误率，并将不匹配的记录保存到新的Excel文件中。

从OCR识别后的医疗单据文本中提取日期、医生姓名、患者姓名、诊断和总消费金额，具备文本矫正和关键字识别能力，并以JSON格式输出。

使用Python Tkinter Treeview组件读取Excel文件，实现点击列头排序、添加动态行序号、双击单元格弹窗显示内容以及自定义列宽的功能。

Advanced Python visualization expert supporting Pandas DataFrames and raw data. Handles dynamic single/dual-axis plotting, Chinese font configuration, Tensor conversion, and safe file saving without display.

Integrates the CosineAnnealingLR learning rate scheduler into the existing training pipeline configuration, allowing dynamic learning rate adjustment based on cosine annealing strategy.

实现用于Vision Transformer的UEP模块，包含特定的卷积层顺序（低维投影、并行卷积、中间处理、高维投影）、残差连接和维度变换，并将其与ViT编码器块进行并行处理集成。

修改多模态视觉Transformer模型，构建双分支架构分别处理RGB和Event数据（分别拼接模板与搜索区域），并输出独立特征以支持双Head处理。

使用Hugging Face Transformers和sacrebleu库，对预训练的机器翻译模型进行评估。支持从制表符分隔的CSV文件读取源文本和参考翻译，计算BLEU分数，并提供Seq2SeqTrainer的compute_metrics函数实现。

读取股票5分钟K线CSV文件，按日期分组，计算基于昨日收盘价的百分比变化，过滤无效数据（长度不一致、NaN、Inf），进行Z-score归一化，并使用TimeSeriesKMeans（DTW/SoftDTW度量）进行聚类。

Parses a URL string to extract query parameters into a key-value dictionary structure without using std::unordered_map.

Design and implement a centralized rate limiting solution for a multi-pod Kubernetes application where API limits are loaded from an external file (e.g., Excel) and synchronized via Redis.

Edits and rewrites sci-fi or fantasy text (novels or game lore) into an epic, high-fantasy style, enhancing vocabulary and tone while strictly preserving original facts and plot points.

Writes expository essays including an outline, introduction, body paragraphs, and conclusion. Body paragraphs must strictly follow the structure: topic sentence, elaboration, examples, and concluding sentence.

Identify and list relevant key terms, concepts, and phrases from provided text blocks.

Create a Python script using regular expressions to extract all positive integers and floating-point numbers from a list of strings, ensuring hyphens are treated as separators and handling optional currency symbols.

Extracts structured CLAIM tuples from HTML tables by distinguishing context vectors from scientific measures, utilizing captions and context for accuracy.

Corrects grammar and vocabulary in IELTS essays, explains errors, suggests synonyms, and enforces structural requirements (e.g., 4-paragraph Task 1) while strictly preserving the user's original wording and structure unless explicitly requested otherwise.

Expands short marketing texts into longer Instagram posts, ensuring emojis are placed at the beginning of each paragraph.

Runs a slow-paced interactive story in second person present tense, pausing after short segments to provide plot suggestions or solicit the user's next action and revisions.

Implement a PyTorch Transformer model using nn.Transformer without manual weight initialization, and a text-to-tensor conversion function for a fixed 8-bit character vocabulary without external libraries.

Solves linear algebraic equations for x and ensures the final answer is fully simplified, preferring exact fractional forms over decimals.

Writes paragraph-based evaluations and reflections for project reports, strictly following a user-provided list of guiding questions or points as the structure.

Performs the fictional 'taphao' mathematical operation defined by the user, where A taphao B equals A multiplied by the sum of the digits of B.

Implements a computer vision pipeline to summarize videos by detecting and tracking multiple objects, selecting only frames containing motion.

Process an array of objects containing timestamps to count occurrences per hour and day, then export the aggregated counts to a CSV file.

Explains or rewrites SQL concepts for beginners using concise language. Ensures SQL examples are logically organized and placed immediately below the relevant explanation paragraphs.

Provides answers to statistics questions in both English and Chinese.

Calculates order quantity using (Balance * Leverage) / Price and integrates it into a Python trading bot loop.

Implement a high score tracking system for Codesters games using CSV files. This involves reading existing scores, appending new scores, sorting them, writing back to the file, and displaying the results on the stage.

Sorts numbers in ascending order based on their column position within a dataset of rows.

Configure a Pydantic BaseSettings class to load environment variables from a .env file, ensuring correct relative or absolute path resolution.

Calculates the number of whole shares to buy at multiple price levels to approximate a target average price, ensuring shares are bought at every level and the total cost stays within budget.

Transforms lists of step-by-step instructions into cohesive paragraphs written in the past tense, strictly avoiding first-person pronouns to maintain an objective, passive voice.

Executes binary classification using DNN and CNN models, with and without CHAID feature selection, using a rolling time-series training window. Handles missing data via mean imputation and outputs a CSV with appended prediction columns.

Provides definitions, mathematical expressions, and simple examples for terms strictly within the context of Fluid Mechanics and Thermodynamics.

Summarize entire books of scripture with a minimum length of six full paragraphs and a descriptive, elaborative style.

Perform a comprehensive SEO audit for a product on e-commerce platforms like Lazada. This involves identifying top competitors, extracting pricing and high-volume keywords, comparing them with the user's product to find gaps, and suggesting title optimizations for better search ranking.

Calculates the number of disorder pairs (inversions) in a list where an element at a lower index is greater than an element at a higher index. Prioritizes efficient algorithms (O(n log n)) suitable for large datasets.

Analyzes accounting data in Excel to calculate days inventory in stock and goodwill ratios, grouped by business sectors derived from SIC codes, including summary statistics and plotting.

Expands a provided topic, sentence, or short text into a single paragraph of exactly 54 words, maintaining the original meaning and context.

Extracts embedding vectors from the layer immediately preceding the Softmax layer of a pre-trained model (e.g., Inception-V3, ResNet50) and saves them in a dictionary where the key is the embedding vector and the value is the corresponding label.

Provides a bash script to automatically identify the Steam ID3 of the currently logged-in user on Linux without using APIs, handling multiple accounts and excluding the default '0' folder.

Adopt a Generation Z persona and speaking style during interactions, using casual language, slang, and abbreviations typical of this demographic.

Runs a second-person interactive story game tracking bladder and bowel levels with graphic detail. The narrative progresses slowly based on user actions, with explicit status updates and descriptive elimination events.

Roleplay as Natalie, a 25-year-old MFA graduate from New York, to explain novel paragraphs and provide entertainment through literary trivia or mini-games.

Transforms a single-column matrix into a multi-column matrix by splitting data into 24-row chunks, appends rows for average and max index, and calculates the most frequent max index.