Our Work

Dataset Description This dataset contains messages identified as instances of political othering in Brazilian public WhatsApp groups. It was developed as part of the study “Us vs. Them: A Datase…

Dataset Description This dataset contains messages identified as instances of political othering in Brazilian public WhatsApp groups. It was developed as part of the study “Us vs. Them: A Dataset of Political Othering in Brazilian WhatsApp Discourse”. The dataset was constructed using a combination of manual annotation and large language model-assisted labeling with systematic human validation. Each record includes the anonymized message text and the corresponding annotation fields used in the study. Personally identifiable information and other potentially sensitive information were removed or anonymized before release. Due to the sensitive nature of the data, the dataset files are not publicly available for download. Access may be granted to researchers upon reasonable request. Requests should include the applicant’s institutional affiliation, the purpose of the research, the intended use of the data, and a commitment not to attempt to identify users or redistribute the dataset.  Researchers using this dataset must cite the associated paper: Daniel Kansaon et al. “Us vs. Them: A Dataset of Political Othering in Brazilian WhatsApp Discourse.” In: Proceedings of the ACM Conference on Hypertext and Social Media (HT ’26), 2026. By requesting access, users agree to use the dataset exclusively for academic and non-commercial research, to comply with applicable ethical and data-protection requirements, and not to redistribute the data or publish identifiable message content.

Code and derived results for a study of whether the chemical transformations thatgenerate activity cliffs on one protein kinase predict cliffs on another. Matched molecular pairs were constructed from…

Code and derived results for a study of whether the chemical transformations thatgenerate activity cliffs on one protein kinase predict cliffs on another. Matched molecular pairs were constructed from measured Ki and Kd binding affinitiesretrieved from ChEMBL (release 37) for 20 human protein kinases, and activity-cliffprediction was evaluated by leave-one-kinase-out cross-validation against explicitprevalence and transformation-frequency baselines. Within-kinase performance wasadditionally assessed under five partitioning schemes of increasing stringency(random pair-level, and grouped by transformation, constant context, Murcko scaffold,and connected molecular component) to quantify the effect of information leakage. The repository contains the end-to-end pipeline, the figure and table generators, thethirteen additional analyses reported in the manuscript (assay-type separation,matched-pair geometry, leakage-controlled validation, correlation robustness, chemicaloverlap, feature ablation, clustered bootstrap, recurrence modelling, thresholdsensitivity, and transformation inventory), and the derived numerical results. All values derive from measured bioactivity data. No simulated, predicted, or imputedbioactivities are included. Analyses run on commodity hardware without a GPU.

initial commit

INFRA-COMPASS is an innovative software tool utilizing Large Language Models (LLMs) to systematically compile and update an inventory of state and local codes and ordinances related to energy infrastr…

INFRA-COMPASS is an innovative software tool utilizing Large Language Models (LLMs) to systematically compile and update an inventory of state and local codes and ordinances related to energy infrastructure. Given the multitude of jurisdictions involved, automating the inventory process through AI is crucial for monitoring and understanding the implementation of these ordinances. The inventory aids in improving and expediting the permitting process in several ways. It enhances transparency for stakeholders, including energy developers, government officials, and communities, by detailing existing siting and permitting requirements. It also facilitates comparisons of regulatory variations across jurisdictions and over time, linking these to permit timelines, technical assistance efforts, and deployment rates to identify best practices. INFRA-COMPASS also provides insights for jurisdictions revising their rules and helps developers conduct comparative analyses to better assess energy development opportunities. Furthermore, INFRA-COMPASS improves the accuracy of energy modeling tools by ensuring they accurately reflect current policies.

Chickpea (Cicer arietinum L.) is one of the most important pulse crops cultivated in India due to its high nutritional value, ability to fix atmospheric nitrogen, and adaptability to diverse agro-clim…

Chickpea (Cicer arietinum L.) is one of the most important pulse crops cultivated in India due to its high nutritional value, ability to fix atmospheric nitrogen, and adaptability to diverse agro-climatic conditions. However, declining soil fertility, low phosphorus availability, and excessive dependence on chemical fertilizers have emerged as major constraints to sustainable chickpea production. Liquid bio-fertilizers containing beneficial microorganisms such as Rhizobium and phosphorus-solubilizing bacteria (PSB) offer an eco-friendly and cost-effective alternative for enhancing nutrient availability, improving soil health, and increasing crop productivity. Keeping these aspects in view, a field experiment entitled “Influence of Nitrogen- and Phosphorus-Based Liquid Biofertilizers on Growth, Yield, and Nutrient Uptake of Chickpea (Cicer arietinum L.)” was conducted during the rabi season of 2022–23 at the Department of Agronomy, School of Agricultural Sciences, Career Point University, Kota, Rajasthan, India. The experiment was laid out in a Factorial Randomized Block Design (FRBD) with three replications, comprising sixteen treatment combinations involving four fertility levels [control (F₀), 10 kg N + 30 kg P₂O₅ ha⁻¹ (F₁), 15 kg N + 40 kg P₂O₅ ha⁻¹ (F₂), and 20 kg N + 50 kg P₂O₅ ha⁻¹ (F₃)] and four liquid biofertilizer treatments [control (LB₀), liquid Rhizobium inoculation (LB₁), liquid PSB inoculation (LB₂), and combined liquid Rhizobium + PSB inoculation (LB₃)]. Chickpea variety GNG-1958 was sown at a spacing of 30 × 10 cm using a seed rate of 80 kg ha⁻¹. Growth parameters, nodulation, yield attributes, seed and haulm yield, nutrient uptake, protein content, chlorophyll content, soil fertility status, and economic returns were recorded and analyzed. The experimental soil was clay loam in texture, slightly alkaline in reaction (pH 8.03), low in available nitrogen, medium in available phosphorus, and high in available potassium. The study aimed to evaluate the influence of integrated nutrient management through nitrogen, phosphorus, and liquid bio-fertilizers on crop growth, productivity, nutrient uptake, and profitability. It was hypothesized that the combined application of recommended nitrogen and phosphorus levels with liquid Rhizobium and PSB inoculation would enhance nodulation, nutrient availability, physiological efficiency, and yield while improving soil fertility and reducing dependence on chemical fertilizers. The investigation provides valuable insights into sustainable nutrient management strategies for chickpea cultivation under the humid south-eastern plain region of Rajasthan. The findings are expected to support the development of economically viable and environmentally sustainable fertilizer recommendations by integrating inorganic fertilizers with liquid bio-fertilizers, thereby contributing to improved productivity, soil health, and long-term agricultural sustainability.

wisp-science v0.22.0 这是继 v0.21.0 之后的功能与稳定性版本,重点把跨 SSH 环境的数据移动纳入可审计、可取消的 Run 流程,并改善会话排队、置顶、分组和动态 Agent 工作流。远程命令失败现在会作为可纠正的探索结果反馈给 Agent,不再把一次普通命令错误误判为 SSH 认证失败;会话侧则新增可编辑队列、任务完成通知跳转和更紧凑的分组移动菜单。 下载 macO…

wisp-science v0.22.0 这是继 v0.21.0 之后的功能与稳定性版本,重点把跨 SSH 环境的数据移动纳入可审计、可取消的 Run 流程,并改善会话排队、置顶、分组和动态 Agent 工作流。远程命令失败现在会作为可纠正的探索结果反馈给 Agent,不再把一次普通命令错误误判为 SSH 认证失败;会话侧则新增可编辑队列、任务完成通知跳转和更紧凑的分组移动菜单。 下载 macOS(.dmg,已签名 + 公证):区分 Apple Silicon(aarch64)与 Intel(x86_64),请下载对应架构。 Windows 安装包(MSI / NSIS):尚未签名,可能触发 SmartScreen,可选择「仍要运行」。 见下方 Assets。 新增功能 结构化 SSH 数据传输 服务器之间直接传输:新增需明确批准的 transfer_between_contexts。可在两个已配置 SSH context 之间直接传输,优先使用 rsync,并在不可用时回退到 scp。 独立凭据的本地中转:源端和目标端无法直接连通时,可通过本机 relay;两端各自使用已保存的连接配置,不要求嵌套 SSH 或共享账号凭据。 定向托管信任:configure_ssh_trust 为明确的 A→B 方向生成专用 Ed25519 信任边,不把私钥复制到本机或 SQLite。 SSH 下载到本地:远程文件可传到用户指定的精确本地新路径。传输使用同卷 staging、拒绝覆盖现有目标,失败或取消后清理不完整数据。 持久化传输 Run:传输进度、吞吐、取消、超时与结果进入现有 Run 生命周期;Agent 通过一次 monitor_run 等待,而不是反复登录或轮询。 防止绕过控制面:run_in_context 会拒绝嵌套 ssh/scp/rsync-over-ssh,并引导 Agent 使用结构化工具。 Agent 工作流、MCP 与研究追溯 动态 Agent 草稿导入/导出:临时 Agent 工作流可用 JSON 保存和恢复,复用当前 proposal 格式;解析失败会显示明确错误。 让模型检查工作流:可把同一份结构化草稿连同审查请求送入主对话,在创建任务前检查目标、依赖、权限和执行器配置。 接管后保留工作流视图:接管 delegated child session 后,Agents 面板、根工作流与任务卡片继续可见;父会话内容没有被截断或删除。 MCP App 状态进入模型上下文:Host 支持标准 updateModelContext 文本能力;应用的最新选择或工作台状态会进入下一轮 native/ACP 请求,关闭或替换 App 后清除。 技能与 MCP 调用更易识别:对话中的工具卡片与折叠步骤会显示「技能」或「MCP」徽章及清理后的真实名称,普通内置工具保持原样。 复制 provenance 代码:Artifact 查看器的 Code 页可直接复制生成图像时记录的只读源代码,并复用现有 Copied 提示。 会话、项目与浏览器体验 可管理的后续消息队列:运行中的会话默认将新消息排队;每条消息可编辑、取消、上移、下移,也可选择插入当前 native Agent 回合。发送菜单继续提供"中断并替换"。 对话置顶:重要会话可置顶到侧栏独立区域并从原分组中抽出,取消置顶后回到原有日期或文件夹分组。 更紧凑的移动菜单:会话菜单只保留一个"移动到"入口;悬停或激活后再选择"未分组"或目标分组,长列表可独立滚动并自动避开窗口边缘。 通知跳转到对应会话:后台任务完成通知使应用窗口重新获得焦点时,会一次性打开产生通知的项目与会话。 标准研究目录改为可选:新建项目可主动勾选标准 layout;默认关闭,不再向已有仓库或不需要该约定的项目写入空目录。约定通过可见、可编辑的 Agent Context 保存,并包含 remote/<server>/ 来源目录建议。 真实浏览器可创建标签页:新增需批准的 web_open_tab;即使当前 profile 只有 chrome://newtab,Agent 也能先打开一个 HTTP(S) 页面,再使用既有扫描与执行工具。 修复内容 远程探索失败语义:失败、超时、取消或 lost 的 waited Run 会作为失败工具结果返回,计划步骤保持未完成;普通远程命令和临时传输错误仍允许 Agent 修正,只有认证拒绝和 host-key 信任失败会阻止再次自动登录。 使用真实解释器路径:环境探测会展示并要求使用实际解释器路径,例如 /usr/bin/python3,避免探测成功后又调用不存在的 python。 流式截断不再伪装成完成:OpenAI/Anthropic 兼容流若没有终止标记且并非用户 Stop,会报告明确的 incomplete 错误,提示网络、relay 或同 key 并发限制,而不是把半句话显示成完整回答。 完成回复即时显示:attempt_completion 的最终回答现在追加在本轮工具调用之后,不再复用工具前的空白 assistant 占位,因此无需重开会话即可看到完整回复。 保持 Rust 1.88 构建兼容:UTF-8 安全截断不再调用较新工具链才稳定的标准库 API,通知依赖也锁定在支持项目最低 Rust 版本的版本。 旧数据库置顶迁移:新增幂等 0023_session_pinned 修复已记录旧迁移但缺少 frames.pinned 的数据库,避免 list_sessions_page 报错并导致整个会话侧栏为空。 语言不再在重开设置时回退:当前会话中的 locale 保持为唯一 live 状态;重新打开设置或管理模型不会用磁盘旧值覆盖界面语言。 调试请求更完整:导出调试请求时保留最新、provider 无关的完整模型请求,包括运行时注入内容。 兼容性与升级说明 可直接覆盖安装 v0.21.0;项目、会话、模型、连接、ExecutionContext、Run、插件与 Agent workflow 数据会继续保留。 缺少 frames.pinned 的旧数据库会在启动时通过幂等迁移自动修复,无需手工执行 SQL。 标准研究目录开关只影响新建项目;现有工作区不会被自动移动或重排。 排队中的后续消息目前只保存在内存中,应用退出或崩溃后不会恢复;已开始或已持久化的正式对话不受影响。 MCP App 的 live model context 属于当前 session runtime,不独立持久化;恢复后的 App 应在初始化时重新发布当前状态。 已知限制 本地 relay 需要约等于源数据大小的可用磁盘空间;scp 传输不可续传,递归 scp 会跟随符号链接。 SSH→本地下载目前使用 scp,不支持断点续传;移除已配置的定向信任边仍需手工操作。 流被截断后不会自动重试,因为已显示并持久化的文本 delta 可能重复;请在错误提示后重新发送或继续。 ACP 协议不支持 native mid-turn guidance,因此 ACP 会话不显示"插队到本回合"。 Windows 安装包仍未签名,可能触发 SmartScreen。 需自备模型与第三方科研服务的 API Key;对应费用、配额与可用性由用户及服务商负责。 测试与交付 Rust workspace、Rust 1.88 / stable、WASM UI 编译和 Playwright 桌面 UI 回归测试在主分支通过。 新增覆盖结构化 SSH 直传/relay/下载/取消/清理、Run 失败语义、动态工作流 JSON round-trip、MCP App context、会话队列与置顶迁移、流式截断、完成回复定位、分组子菜单和新项目 layout 开关。 v0.21.0..v0.22.0 包含 22 个非合并功能与修复提交,涉及 48 个文件。 推送 v0.22.0 标签后,GitHub Actions 会构建 macOS Apple Silicon / Intel DMG 与 Windows MSI / NSIS,并附加到本 Release。 反馈 欢迎在 GitHub Issues 反馈问题。SSH 传输问题请说明源/目标 context、所选 route、Run 状态和脱敏后的错误;队列、Agent workflow 或 MCP App 问题请附上复现步骤及相关会话状态。API key、token、密码、Cookie、SSH 私钥和未脱敏科研数据请勿上传。 English: wisp-science v0.22.0 adds structured, auditable transfers across SSH contexts and from SSH to local storage; editable queued follow-ups and pinned conversations; reusable dynamic Agent workflow drafts; and clearer MCP, Skill, and provenance surfaces. Remote command failures now remain correctable exploration instead of being confused with authentication failures, while cut model streams surface an explicit incomplete-response error. What's Changed Transfer files directly between configured SSH contexts with rsync/scp, or relay through the local machine with each host's own credentials. Configure an explicit directed Ed25519 trust edge without copying private keys into local storage or SQLite. Download SSH files to an exact new local path through a persisted transfer Run with staging, cancellation, timeout, and cleanup. Return failed or cancelled waited Runs as failed tool results, preserve unfinished plan steps, and use the exact interpreter path found by environment probes. Queue follow-up messages by default while a conversation runs; edit, cancel, reorder, cut in, or interrupt-and-replace them. Pin conversations above normal grouping, and repair existing databases that recorded the old migration without the frames.pinned column. Import/export dynamic Agent workflow drafts as JSON and send the same proposal to the model for review before creation. Keep root workflow cards visible after taking over a delegated child session. Inject the latest standard MCP App updateModelContext state into the next native or ACP turn, and label MCP/Skill calls clearly in the transcript. Copy recorded artifact provenance code from the viewer. Move folder destinations into a viewport-aware submenu, jump from completed-task notifications to their session, and make the standard research layout opt-in for new projects. Open the first scriptable HTTP(S) browser tab with the approved web_open_tab tool. Detect OpenAI/Anthropic-compatible streams that end without a terminal marker, preserve live locale when reopening Settings, and export the exact latest provider-agnostic debug request. Render the live attempt_completion answer after its tool rows so the final response is visible immediately without reopening the conversation. Preserve the Rust 1.88 build floor by using a stable UTF-8 truncation path and locking the notification dependency to a compatible release. Compatibility: Installs over v0.21.0 in place. The missing pinned-session column is repaired automatically. Queued follow-ups and live MCP App model context remain runtime-only. The standard research layout is now opt-in for newly created projects and does not reorganize existing workspaces. Full Changelog: https://github.com/xuzhougeng/wisp-science/compare/v0.21.0...v0.22.0

This record contains the PDF of the manuscript “Rational Trace Discontinuities and Irrational Continuity for Kwapisz-Like Rotation Sets.” The paper studies an explicit family of rotation s…

This record contains the PDF of the manuscript “Rational Trace Discontinuities and Irrational Continuity for Kwapisz-Like Rotation Sets.” The paper studies an explicit family of rotation sets associated with torus dynamics. It determines the exact one-sided Hausdorff traces at every rational parameter, proves continuity at every irrational parameter, and establishes strict jump discontinuities at every rational parameter. It also develops quantitative Hausdorff and area separation bounds, Diophantine regularity results, Liouville irregularity, and continued-fraction structure in the rational trace polygons. This version is deposited for open access, citation, and long-term preservation.

A Pesquisa Nacional de Saúde (PNS) consolidou-se como o principal inquérito domiciliar de base populacional voltado à investigação das condições de sa&…

A Pesquisa Nacional de Saúde (PNS) consolidou-se como o principal inquérito domiciliar de base populacional voltado à investigação das condições de saúde da população brasileira. Este capítulo tem por objetivo sistematizar, por meio de revisão integrativa da literatura científica brasileira publicada entre 2020 e 2026, os principais indicadores epidemiológicos revelados pela edição de 2019 da pesquisa, em comparação com a edição de 2013. Foram consultadas as bases SciELO, BVS, LILACS, PubMed/MEDLINE e Google Acadêmico, além de publicações técnicas oficiais do IBGE e do Ministério da Saúde. Os resultados evidenciam aumento expressivo da prevalência de obesidade, de doenças crônicas não transmissíveis e de depressão autorreferida, além da persistência de desigualdades regionais, sociais e raciais no acesso aos serviços de saúde. Discute-se, por fim, o papel da PNS como instrumento estratégico de vigilância epidemiológica, considerando o início recente da coleta de dados de sua terceira edição, a PNS 2026.