Program Synthesis

A new large dataset with over 100,000 examples consisting of Java classes from online code repositories, and develop a new encoder-decoder architecture that models the interaction between the method documentation and the class environment. Source: Mapping Language to Code in Programmatic Context

DeepFix consists of a program repair dataset (fix compiler errors in C programs). It enables research around automatically fixing programming errors using deep learning.

HumanEvalPack is an extension of OpenAI's HumanEval to cover 6 total languages across 3 tasks. The evaluation suite is fully created by humans.

The ManySStuBs4J corpus is a collection of simple fixes to Java bugs, designed for evaluating program repair techniques. We collect all bug-fixing changes using the SZZ heuristic, and then filter these to obtain a data set of small bug fix changes. These are single statement fixes, classified where possible into one of 16 syntactic templates which we call SStuBs. The dataset contains simple statement bugs mined from open-source Java projects hosted in GitHub. There are two variants of the dataset. One mined from the 100 Java Maven Projects and one mined from the top 1000 Java Projects. The dataest contains 153,652 single statement bugfix changes mined from 1,000 popular open-source Java projects, annotated by whether they match any of a set of 16 bug templates, inspired by state-of-the-art program repair techniques.

SketchGraphs is a dataset of 15 million sketches extracted from real-world CAD models intended to facilitate research in both ML-aided design and geometric program induction. Each sketch is represented as a geometric constraint graph where edges denote designer-imposed geometric relationships between primitives, the nodes of the graph. Source: https://github.com/PrincetonLIPS/SketchGraphs Image Source: https://github.com/PrincetonLIPS/SketchGraphs

Pseudocode-to-Code (SPoC) is a program synthesis dataset, containing 18,356 programs with human-authored pseudocode and test cases. Image source: https://sumith1896.github.io/spoc/

xCodeEval is one of the largest executable multilingual multitask benchmarks consisting of 17 programming languages with execution-level parallelism. It features a total of seven tasks involving code understanding, generation, translation, and retrieval, and it employs an execution-based evaluation instead of traditional lexical approaches. It also provides a test-case-based multilingual code execution engine, ExecEval that supports all the programming languages in xCodeEval.

Defects4J is a collection of reproducible bugs and a supporting infrastructure with the goal of advancing software engineering research. Defects4J contains 835 bugs (plus 29 deprecated bugs) from the following open-source projects: | Identifier | Project name | Number of active bugs | Active bug ids | Deprecated bug ids (*) | |-----------------|----------------------------|----------------------:|---------------------|-------------------------| | Chart | jfreechart | 26 | 1-26 | None | | Cli | commons-cli | 39 | 1-5,7-40 | 6 | | Closure | closure-compiler | 174 | 1-62,64-92,94-176 | 63,93 | | Codec | commons-codec | 18 | 1-18 | None | | Collections | commons-collections | 4 | 25-28 | 1-24 | | Compress | commons-compress | 47 | 1-47 | None | | Csv | commons-csv | 16 | 1-16 | None | | Gson | gson | 18 | 1-18 | None | | JacksonCore | jackson-core | 26 | 1-26 | None | | JacksonDatabind | jackson-databind | 112 | 1-112 | None | | JacksonXml | jackson-dataformat-xml | 6 | 1-6 | None | | Jsoup | jsoup | 93 | 1-93 | None | | JxPath | commons-jxpath | 22 | 1-22 | None | | Lang | commons-lang | 64 | 1,3-65 | 2 | | Math | commons-math | 106 | 1-106 | None | | Mockito | mockito | 38 | 1-38 | None | | Time | joda-time | 26 | 1-20,22-27 | 21 |

Five curated datasets of one-liner commits from open-source projects. In total, they are composed of 58069 one-liner commits.

DOI Type Inference dataset for TypeScript. Click on DOI tag for dataset files.

Repair AST parse (syntax) errors in Python code

A massive, deduplicated corpus of 7.4M Python files from GitHub. Source: Learning and Evaluating Contextual Embedding of Source Code

This paper presents a benchmark data set for condition monitoring of rolling bearings in combination with an extensive description of the corresponding bearing damage, the data set generation by experiments and results of datadriven classifications used as a diagnostic method. The diagnostic method uses the motor current signal of an electromechanical drive system for bearing diagnostic. The advantage of this approach in general is that no additional sensors are required, as current measurements can be performed in existing frequency inverters. This will help to reduce the cost of future condition monitoring systems. A particular novelty of the present approach is the monitoring of damage in external bearings which are installed in the drive system but outside the electric motor. Nevertheless, the motor current signal is used as input for the detection of the damage. Moreover, a wide distribution of bearing damage is considered for the benchmark data set. The results of the classifications show that the motor current signal can be used to identify and classify bearing damage within the drive system. However, the classification accuracy is still low compared to classifications based on vibration signals. Further, dependency on properties of those bearing damage that were used for the generation of training data are observed, because training with data of artificially generated and real bearing damages lead to different accuracies. Altogether a verified and systematically generated data set is presented and published online for further research

Project: Discrete-Time Modeling of Interturn Short Circuits in Interior PMSMs Authors: Lukas Zezula, Matus Kozovsky, Ludek Buchta and Petr Blaha (corresponding author: Lukas Zezula, e-mail: lukas.zezula@ceitec.vutbr.cz) Affiliation: CEITEC - Central European Institute of Technology, Brno University of Technology, 612 00 Brno, Czech Republic Files included: ├── Data_diverse_FI/ # Measured and DTM-generated data for an ISC occurring with diverse fault indicators stored as timeseries objects in .mat files │ ├── Rsc_changes_3turns_1Nm.mat # Rsc changes at 1900 rad/s with 1 Nm load, sgm=3/25, Rsc decreasing in 14.4 + {442, 47.0, 5.62, 1.74} mOhm │ ├── Rsc_changes_3turns_3Nm.mat # Rsc changes at 1900 rad/s with 3 Nm load, sgm=3/25, Rsc decreasing in 14.4 + {442, 47.0, 5.62, 1.74} mOhm │ ├── Rsc_changes_6turns_1Nm.mat # Rsc changes at 1900 rad/s with 1 Nm load, sgm=6/25, Rsc decreasing in 14.4 + {442, 47.0, 5.62, 1.74} mOhm │ └── Rsc_changes_6turns_3Nm.mat # Rsc changes at 1900 rad/s with 3 Nm load, sgm=6/25, Rsc decreasing in 14.4 + {442, 47.0, 5.62, 1.74} mOhm │ ├── Data_diverse_OC/ # Measured and DTM-generated data for an ISC occurring at diverse operating conditions stored as timeseries objects in .mat files │ ├── load_transient.mat # Load transient at 1400 rad/s from 1 to 3 Nm with a 50 Nm/s rate, Rsc = (1.74 + 14.4) mOhm, sgm = 10/25 │ ├── velocity_transient.mat # Velocity transient from 1200 to 1600 rad/s with a 10000 rad/s^2 rate and 2 Nm load, Rsc = (1.74 + 14.4) mOhm, sgm = 10/25 │ └── steady_state.mat # Velocity steady state at 1400 rad/s with a 1 Nm load, Rsc = (1.74 + 14.4) mOhm, sgm = 10/25 │ ├── Models/ # Models prepared in MATLAB Simulink, MATLAB 2024b │ ├── codegen_mod_eul.slx # Single-precision implementation of the forward Euler DTM for code generation │ ├── codegen_mod_new.slx # Single-precision implementation of the derived DTM for code generation │ └── models_comparison.slx # Comparing the DTMs with the continuous-time model │ └── README.txt # This file Data in .mat files: idq_eul ... Direct and quadrature axis currents (respectively) predicted by the forward Euler DTM idq_meas ... Direct and quadrature axis currents (respectively) transformed from the measured stator currents idq_new ... Direct and quadrature axis currents (respectively) predicted by the derived DTM if_eul ... Fault current predicted by the forward Euler DTM if_meas ... Fault current measured in the FIU if_new ... Fault current predicted by the derived DTM Rsc ... Emulated short circuit resistance sgm ... Emulated fault severity SinCos ... Angle tracking observer outputs: sine and cosine (respectively) of the electrical angle Te_eul ... Electromagnetic torque predicted by the forward Euler DTM Te_meas ... Electromagnetic torque calculated from the post-processed measured torque Te_new ... Electromagnetic torque predicted by the derived DTM udq ... Direct and quadrature axis voltages (respectively) generated by the control algorithm we ... Angle tracking observer outputs: estimated electrical angular velocity

https://arxiv.org/abs/2106.06086

This data comprises processed weather, soil, yield, and cultivation area for corn yield prediction in Sub-Sahara Africa, with emphasis on Nigeria. The data was collected to design a corn yield prediction model to help smallholder farmers make smart farming decisions. However, the data can serve several other purposes through analysis and interpretation. The reference study region in Africa is Nigeria. The focuses on corn crop because there are over 211.4 million people, of which a large percentage of the population are smallholder farmers. Nigeria [9.0820° N, 8.6753° E] is within an arable land area of 34 million hectares located on the west coast of Africa. The region comprises of 36 states with the most and least number of districts being 214 and 10, respectively. For each state, the environment data are collected as follows. Grid map climate data – This data spans spatial resolutions between ~1 km2 to ~340 km2 from the high spatial resolution WorldClim global climate database22. Each grid point on the map is monthly data from January to December between 1970 and 2000 years and records 8 climate variables. The variables are average temperature C0, minimum temperature C0, maximum temperature C0, precipitation (mm), solar radiation (kJ m^(-2) day(-1), wind speed (m s(-1)), and water vapor (kPa) taken at 30 seconds (s), 2.5 minutes m, 5 m, and 10 m. Grid map soil data – This data is obtained from 250 minutes of spatial resolution AfSIS soil data23 from year 1960 to 2012. The variables are wet soil bulk density, dry bulk density (kg dm-3), clay percentage of plant available water content, hydraulic conductivity, the upper limit of plant available water content, the lower limit of, organic matter percentage, pH, sand percentage (g 100 g-1), silt percentage (g 100 g-1) and, clay percentage (g 100 g-1), and saturated volumetric water content variables measured at depths 0–5, 5–10, 10–15, 15–30, 30–45, 45–60, 60–80, 80–100, and 100–120 measured in centimeters (cm). Corn yield data – This data is available on Kneoma Corporation website24. It ranged from years 1995 to 2006 and consisted of a corn yield of 1000 metric tonnes and a cultivation area of 1000 hectares. Geolocation coordinates (latitude and longitude) – The geolocation of each of the 36 states with their districts is sampled from Google Maps. The output feds into the Esri-ArcGIS 2.5, a professional geographical software, for extracting the point-cloud values of each environmental variable (weather and soil) at specific geolocation of the 36 states of Nigeria. Other Descriptions: Data type - Continous and Categorical Dataset Characteristics - Tabular Associated Tasks - Regression Feature Type - Real Number of Instances - 1828 Number of Features: 12

The dataset contains standard contexts of the lattices of all atomic lattices in the Concept Explorer format.

The dataset contains more than 100k code patch pairs extracted from open source projects on GitHub. Each pair comes with the erroneous and the fixed version of the corresponding code snippet. Instead of the whole file, the code snippets are extracted to focus on the problematic region (error line + other lines around it). For each sample, the repository name, the commit id, and the file names are provided so that one can access the complete files in case of interest. The dataset only has JavaScript programs and the error are detected by the popular static code analyzer ESLint. The dataset can be used in the fields of: program repair, code generation, bug finding, transfer learning and many more fields related to machine learning for code