Explore Projects

The task is to make cpimport support parquet files as input or design a general framework to add new input formats. Considering cpimport finish its workflow by two stages, read and parse, so I plan to solve this problem from these two stages. For read stage, I will use some external library to help finish reading like `parquet-cpp`, `arrow` and so on. By using parquet-cpp, I can directly read parquet file and by using `arrow::table` I can store data in table format. Later, in parse stage, I plan to use table format data and parse it to insert data into target table. After finishing the task, cpimport can handle parquet file input or other poplular formats like arrow, avro and successfully insert data into the database.

Kyverno currently supports cluster-wide CEL-based policy types (ValidatingPolicy, MutatingPolicy, ImageValidatingPolicy, GeneratingPolicy, and DeletingPolicy), but lacks namespaced versions. This creates challenges for namespace owners who need to manage policies within their namespaces without requiring cluster-wide permissions. The goal of this project is to implement namespaced versions of all five CEL-based policy types to provide better RBAC control, security, and lifecycle management for namespace-scoped policy management. Expected Outcome: 1. Implement namespaced versions of all five CEL-based policy types 2. Ensure proper RBAC integration for namespace-scoped policy management 3. Create comprehensive documentation and examples for namespaced policies 4. Write tests to validate namespaced policy functionality

Virtio-net device uses Control Virtqueue(CVQ) for changing device parameters. For some devices, such as vDPA, its CVQ is passed from Guest’s driver directly to Host driver, which makes it difficult for QEMU to track the status of these devices. To solve this problem, QEMU introduces Shadow Virtqueue(SVQ) for vdpa device, which shadows the CVQ via QEMU instead of being assigned directly to Guest. During QEMU Live migration, to restore the vdpa device in the destination VM to the state in the source VM, SVQ can use CVQ’s interface and send the state as regular CVQ commands. Due to some CVQ commands for vDPA are missing at present, this project aims to achieve support for a subset of CVQ commands, including VIRTIO_NET_CTRL_RX family commands, VIRTIO_NET_CTRL_GUEST_OFFLOADS_SET command, VIRTIO_NET_CTRL_VLAN_ADD command, VIRTIO_NET_CTRL_MQ_HASH_CONFIG command and VIRTIO_NET_CTRL_MQ_RSS_CONFIG command for vDPA.

This project aims to enhance Wayshot by implementing the newly standardized ext_image_capture_source_v1 Wayland protocol, which provides a standardized method for display and window capture, replacing the current dependency on the non-standard wlr-screencopy protocol. I will develop a new backend that fully utilizes this official protocol while maintaining the existing wlr backend for backwards compatibility, enabling native top-level window capture capabilities and improving user experience. The implementation will include a security notification system utilizing D-Bus to alert users which applications were captured by Wayshot clients, enhancing transparency and user privacy. Additionally, I will integrate the standardized protocol with xdg-desktop-portal-luminous to enable WebRTC streaming capabilities, The deliverables of this project are: a complete solution that includes the new backend, backwards compatibility, security notifications, WebRTC streaming support, and comprehensive documentation.

<p>A large amount of 3D data is generated in High Energy Physics &amp; Astrophysics experiments. To process this data efficiently, one would need state-of-the-art tools. Already a lot of development has been done in processing 2D data with projects like spatial Hadoop and GeoSpark but, there are very few frameworks to process the 3D data. The idea is to follow the footsteps of GeoSpark and provide a way to load, process and analyse 3D data sets economically and efficiently by leveraging the distributed computation functionality of the spark. Spark3D would provide the set of out-of-the-box 3D Spatial RDD (3D SRDD) to partition the data across machines. Ultimately, Spark3D would be available as an open-source library which works with all recent versions of the Spark (2.0+), has user friendly APIs (in Scala, Java and Python), works on top of all major platforms out of the box (HDFS, S3, Cassandra, etc.) and supports all major file formats (CSV, Parquet, JSON, Avro, etc.) including popular scientific file formats such as FITS.</p>

<p>Add Active/Standby HA support for the KubeSlice Controller.</p><p>&nbsp;- Leader election: Only one controller cluster (the “Active” node) holds a distributed lock (e.g., a Lease) and is permitted to write updates to worker clusters or manage Slice configurations.</p><p>&nbsp;- Constant state sync: The Standby cluster continuously mirrors the Active cluster’s state, including relevant KubeSlice CRDs (Slices, ServiceExports, Clusters, etc.).</p><p>&nbsp;- Heartbeating: The Standby cluster monitors the health of the Active cluster.</p><p>&nbsp;- Failover: If the Active cluster fails to renew its lease, the Standby cluster detects the timeout, acquires the lock, promotes itself to Active, and worker clusters are updated to use the new Active controller.</p><p><br></p><p>Expected Outcome: Implement an Active/Standby HA architecture that makes the KubeSlice management plane resilient and disaster-recovery-ready.</p><p><br></p>

Istio extends Kubernetes to establish a programmable, application-aware network. Working with both Kubernetes and traditional workloads, Istio brings standard, universal traffic management, telemetry, and security to complex deployments Istio does not support HTTPs based metric scraping for control plane, gateway, and Envoy sidecar metrics This could have some security related consequences: - An attacker might find some sensitive information that they can use for their advantage. For example, Envoy /stats endpoint can be used to enumerate all upstream services in the cluster. - In theory an attacker could masquerade the metrics endpoint(s) and inject fake data to monitoring systems, in order to e.g. hide an ongoing attack, confuse the system to autoscale up/down etc. It would be nice to protect the metrics endpoints with TLS, using mutual authentication. While this feature is a big one covering multiple components, the easiest component alone is intended to be covered as part of this internship. Expected Outcome: - Implement HTTPS metrics for ztunnel component - Add unit tests and integration tests for the feature - Add documentation for the functionality

<p>Optical sensors are unable to penetrate clouds leading to related incorrect reflectance values. Unlike Landsat images, Sentinel 2 datasets do not include thermal and Quality Assessment bands that simplify the detection of clouds avoiding erroneous classification. At the same time, clouds shadows on the ground lead to anomalous reflectance values which have to be taken into account during the image processing. The project aim is the coding of a specific module for GRASS GIS application which implements the procedure developed within my PhD research. The procedure allows to automatically identify clouds and their shadows in Sentinel 2 images applying some rules on reflectance values (values thresholds, comparisons between bands, etc.). These have been defined starting from rules found in literature and conveniently refined. Then the detection of shadows is improved using an adapted shape index. In order to increase the accuracy of the final results, a control check is implemented. Clouds and shadows are spatially intersected in order to remove misclassified areas. The final outputs are two different vector files (OGR standard formats), one for clouds and one for shadows.</p>

<p>To develop an Android &amp; IOS application to provide clinical decision-support for nurses and doctors delivering essential newborn care interventions during the first day of life. This application provides knowledge, skills, and competencies to nurses and doctors in low/middle-income settings so that they can provide life-saving care to newborns from birth through 24 hours postnatal. App uses DHIS2 as backend &amp; can be setup for any hospital/organization</p> <p>The Goals of this project are</p> <ol> <li>Ability to track multiple babies at a time; </li> <li>Time-stamping births to generate an essential newborn care clock for each baby being tracked in the app;</li> <li>Automated classification of babies health status, based on results from observations and assessments that are conducted during the first 2 hours after birth;</li> <li>Automated advice on management of babies, based on their classification; Reminders/alerts to provide recommended essential newborn care interventions; </li> <li>If particular interventions are not delivered on time, or at all, capture of information regarding why this essential newborn care was late, or missed</li> </ol>

Lerobot is a framework for rapidly developing embodied intelligent robots, which includes classic imitation learning strategies such as ACT and Diffusion Policy, as well as low-cost robot hardware. It is necessary to combine it with dora. Recently, support for the SO-ARM100, a low-cost robotic arm, was added to Dora through its driver. However, there is currently no end-to-end tutorial available for using this arm. Tutorials and documentation are critical for making the project more accessible to new users. The goal of this project is to improve the Python-based tutorial for using the SO-ARM100 robotic arm with Dora-LeRobot, enabling users to easily get started and contribute to the ecosystem. Expected Outcomes: 1.Fix issues and add missing code so that Dora-LeRobot can be run with SO-ARM100 successfully. 2.Writing detailed setup and usage tutorials. 3.Implementing code for calibration, control, training, and inference. 4.Creating video tutorials and improve based on community feedbacks. 5.Creating test plan for user testing to ensure the tutorial is easy to follow. 6.Developing the Rust library for UR5 control. A well-documented, efficient, and effective Rust library that interfaces with UR5 robots through Dora.

The project involves integrating pixel size support into ilastik, an image-processing and classification tool. Pixel size is a metadata item enabling conversion of image pixels into real-world measurements. Whilst pixel size is not currently processed in ilastik image sets, the program is used extensively in scientific applications, and introducing real-world measurements would enable development of several key features that would allow more in-depth quantitative analysis of image data, such as object displacement and velocity in video tracking, and real-world area calculation of segmentations/ volume calculation of carvings. Once the metadata is extracted at the image input end, it is possible to feed it directly into the data pipeline, so it can be picked back up within any relevant workflow. This facilitates feature development, as it then becomes possible to work directly with the metadata values within the workflow. Once the backends of these features have been developed, they will be implemented on the front-end via Qt. The key deliverables of this project are: Implement pixel size metadata into workflows, such that the metadata remains present in project files, output images, and exported datasets; Enable calculating areas of image segmentations; Enable calculating volumes of image carvings; Enable recording of object displacement in video tracking; Enable measurement of object velocity in video tracking; Produce detailed and well-written documentation, and ensure relevant test scripts for each feature are made available. An initial pull request for this proposal can be found at https://github.com/ilastik/ilastik/pull/2995

COCONUT-SVSM is a Secure Virtual Machine Service Module running at a privileged level inside Confidential VMs to provide trusted services (e.g., virtual TPM) independent of the host hypervisor. Currently, the SVSM kernel performs full TLB flushes on every context switch, discarding all TLB entries indiscriminately and causing significant runtime TLB miss overhead. This project enhances COCONUT-SVSM to leverage the x86 PCID (Process Context Identifiers) hardware feature to tag TLB entries per task, preserving them across context switches and enabling selective TLB invalidation. The work is structured in three phases: Phase 1 builds a code-generated Rust crate (x86-cpuid) from the authoritative x86-cpuid-db v2.4 XML database, replacing all scattered, magic-bit-position CPUID checks across the codebase with a type-safe cpu_has(leaf_0x1::ecx::Pcid) API unified across SNP, TDX, and native platforms via a CpuidReader trait. Phase 2 designs and implements a PCID assignment strategy for all virtual pages in the COCONUT kernel, enabling the CPU to tag TLB entries per task(same for all threads of task) and preserve them across context switches. Phase 3 converts global TLB flush operations in into selective, PCID-aware invalidations: using AMD's INVLPGB broadcast instruction to flush a specific PCID across all cores without IPIs. Key Deliverables are x86-cpuid Rust crate with code generator consuming x86-cpuid-db XML, refactored COCONUT-SVSM codebase with all open-coded CPUID checks replaced with new code generated by x86-cpuid Rust crate, PCID allocator, applying the PCIDs in the COCONUT page-tables and selective TLB flush functions (flush_tlb_pcid, flush_tlb_range_pcid) with AMD INVLPGB.

<p>digiKam supports images export to many services. This project aims to add Twitter and OneDrive to digiKam export services.</p>

<p>Implement a function which will receive a text and return tuples of found date substrings and parsed datetime objects for languages currently supported by dateparser.</p>

Develop back and front end support for z/OS in the Flask framework (Python), and work with the team to create the initial JSON source file for open source z/OS software

Develop an autonomous multi-agent system using LangGraph to automatically analyze Hugging Face model architectures and generate the necessary code, tests, and documentation to add support for new models in the Optimum-Intel repository

<p>This project expands Django's parallel test runner to work on Windows and macOS by running the parallel test runner using spawn, and also adds Oracle backend support to using the parallel test runner.</p>

<p>In this project, I will implement a new Linux kernel API to help dumping/writing back the read and write queue of UDP socket. Then I will use that new API to add UDP's read and write queue checkpoint/restore support in CRIU.</p>

<p>gocloud is a standard library for Go that abstracts away differences among multiple cloud providers. The gocloud package is designed to make it easy to provision and work with cloud providers. The proposed project aims to Extending GoCloud provider support for ali cloud, Microsoft Azure and OpenStack.</p>

<p>Pocket Code already implements face and text detection, text recognition, speech recognition and speech synthesization, but these features cannot be used by users with recent Huawei phones. The goal of this project is to support HMS Machine learning and Artificial Intelligence API, where other APIs are not available. So that they too can make use of these features.</p>

With the launch of RocketMQ 5.0 architecture, the RocketMQ Dashboard has not been well adapted to RocketMQ 5.0. There have been issues with Dashboard being unable to create various themes for V5, lack of support for Proxy components, and incorrect master-slave synchronization metrics. Therefore, we need to adapt the RocketMQ Dashboard to the RocketMQ 5.0 architecture in this project to enhance its usability.

The primary goal of this project is to improve the performance of Zulip, with a focus on the "upload" and "import/export" areas. Specifically, I will implement chunk file upload support for large-sized files and enable the direct importing of exports from other services via the web. Additionally, I will work on the images thumbnail part, Which currently places a significant load by downloading the full image.

The major objectives of the project are: 1. Refactoring the Python code to increase efficiency and readability and increase the overall code quality of the entire codebase of the software. 2. To enhance user engagement and support by creating a comprehensive FAQ dataset which is used to train a custom Chatbot. 3. Investigate and discuss different possibilities for multi-GPU model execution, i.e. allow a model to execute (and share memory) across multiple GPUs.