chore: auto-commit (6 archivos)

- CMakeLists.txt
- main.cpp
- data_collectors.cpp
- data_collectors.h
- runner.cpp
- runner.h

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

CMakeLists.txt

@@ -20,6 +20,8 @@ endif()
 add_imgui_app(odr_console
     main.cpp
     data_registry.cpp
+    data_collectors.cpp
+    runner.cpp
 )
 
 target_include_directories(odr_console PRIVATE

data_collectors.cpp

@@ -0,0 +1,77 @@
+#include "data_collectors.h"
+
+#include <filesystem>
+#include <fstream>
+
+namespace fs = std::filesystem;
+
+namespace {
+
+std::string trim(std::string s) {
+    auto issp = [](unsigned char c) { return std::isspace(c); };
+    while (!s.empty() && issp((unsigned char)s.back())) s.pop_back();
+    size_t i = 0;
+    while (i < s.size() && issp((unsigned char)s[i])) ++i;
+    return s.substr(i);
+}
+
+// Strip wrapping quotes if present.
+std::string unquote(std::string s) {
+    if (s.size() >= 2) {
+        char a = s.front(), b = s.back();
+        if ((a == '"' && b == '"') || (a == '\'' && b == '\'')) {
+            return s.substr(1, s.size() - 2);
+        }
+    }
+    return s;
+}
+
+// Parsing minimo: busca lineas top-level `id:`, `name:`, `description:`.
+// Suficiente para metadata de collectors. Si la app necesita mas, anadir yaml-cpp.
+void parse_manifest(const fs::path& yaml_path, Collector& c) {
+    std::ifstream f(yaml_path);
+    if (!f.is_open()) return;
+    std::string line;
+    while (std::getline(f, line)) {
+        std::string t = trim(line);
+        if (t.empty() || t[0] == '#') continue;
+        // Ignora lineas indentadas (sub-keys de listas/dicts).
+        if (line.size() > t.size() && line[0] == ' ') continue;
+
+        auto colon = t.find(':');
+        if (colon == std::string::npos) continue;
+        std::string key = trim(t.substr(0, colon));
+        std::string val = trim(t.substr(colon + 1));
+        val = unquote(val);
+
+        if (key == "id" && !val.empty()) c.id = val;
+        else if (key == "name" && !val.empty()) c.name = val;
+        else if (key == "description" && !val.empty()) c.description = val;
+    }
+}
+
+} // namespace
+
+bool collectors_discover(const std::string& collectors_root,
+                         std::vector<Collector>& out) {
+    out.clear();
+    std::error_code ec;
+    if (!fs::is_directory(collectors_root, ec)) return false;
+
+    for (auto& entry : fs::directory_iterator(collectors_root, ec)) {
+        if (ec) break;
+        if (!entry.is_directory()) continue;
+        Collector c;
+        c.dir = entry.path().string();
+        c.id  = entry.path().filename().string();
+        c.run_py = (entry.path() / "run.py").string();
+        c.has_run = fs::exists(c.run_py);
+
+        fs::path manifest = entry.path() / "manifest.yaml";
+        if (fs::exists(manifest)) parse_manifest(manifest, c);
+        if (c.name.empty()) c.name = c.id;
+
+        out.push_back(std::move(c));
+    }
+    return true;
+}

data_collectors.h

@@ -0,0 +1,18 @@
+#pragma once
+
+#include <string>
+#include <vector>
+
+struct Collector {
+    std::string id;          // basename del directorio (ej. "api_hn_top")
+    std::string name;        // del manifest (fallback id)
+    std::string description; // del manifest
+    std::string dir;         // path absoluto al directorio del collector
+    std::string run_py;      // path absoluto al run.py
+    bool        has_run = false;
+};
+
+// Enumera <collectors_root>/*/manifest.yaml + run.py.
+// Parsing manifest minimo: solo lee `id`, `name:`, `description:` por linea (sin yaml-cpp).
+bool collectors_discover(const std::string& collectors_root,
+                         std::vector<Collector>& out);

main.cpp

@@ -12,19 +12,34 @@
 #include "core/logger.h"
 
 #include "data_registry.h"
+#include "data_collectors.h"
+#include "runner.h"
 
 #include <cstdio>
 #include <cstdlib>
 #include <cstring>
+#include <filesystem>
+#include <sstream>
 #include <string>
 #include <vector>
 
 static OdrRegistry g_registry;
 static std::string g_db_path;
+static std::string g_app_dir;        // dir de odr_console (collectors/, ops db)
+static std::string g_python_exe;     // resolved al arrancar
+static std::string g_registry_root;  // FN_REGISTRY_ROOT
+static std::string g_ops_db_path;    // <app_dir>/operations.db
+
 static char        g_search_buf[256] = "";
 static std::vector<RegistryRow> g_results;
 static int         g_selected = -1;
 
+static std::vector<Collector> g_collectors;
+static int                    g_coll_selected = -1;
+static int                    g_coll_param_limit = 30;
+static std::string            g_last_run_summary; // texto status del ultimo run
+static bool                   g_running = false;
+
 static void do_search() {
     g_results.clear();
     g_selected = -1;
@@ -36,9 +51,126 @@ static void do_search() {
 }
 
 static bool g_show_launcher   = true;
+static bool g_show_collectors = true;
 static bool g_show_jobs       = true;
 static bool g_show_datasets   = true;
 
+static void run_selected_collector() {
+    if (g_coll_selected < 0 ||
+        g_coll_selected >= (int)g_collectors.size()) return;
+    const Collector& c = g_collectors[g_coll_selected];
+    if (!c.has_run) return;
+
+    std::ostringstream ctx;
+    ctx << "{"
+        << "\"ops_db_path\":\"" << g_ops_db_path << "\","
+        << "\"app_dir\":\""     << g_app_dir     << "\","
+        << "\"registry_root\":\"" << g_registry_root << "\","
+        << "\"params\":{\"limit\":" << g_coll_param_limit << "}"
+        << "}";
+
+    std::string tmp_dir = std::string(fn::local_path("runner_tmp"));
+    g_running = true;
+    auto res = odr::run_collector(g_python_exe, c.run_py, tmp_dir, ctx.str());
+    g_running = false;
+
+    std::ostringstream s;
+    s << "exit=" << res.exit_code
+      << "  ms=" << res.duration_ms
+      << "\nstdout: " << res.stdout_str
+      << "\nstderr (last):\n";
+    // Solo ultimas 10 lineas de stderr para no inundar UI.
+    std::vector<std::string> lines;
+    std::istringstream ss(res.stderr_str);
+    std::string ln;
+    while (std::getline(ss, ln)) lines.push_back(ln);
+    int start = (int)lines.size() - 10;
+    if (start < 0) start = 0;
+    for (int i = start; i < (int)lines.size(); ++i) s << lines[i] << "\n";
+    g_last_run_summary = s.str();
+}
+
+static void draw_collectors() {
+    if (!g_show_collectors) return;
+    if (!ImGui::Begin(TI_PLAYER_PLAY " Collectors", &g_show_collectors)) {
+        ImGui::End();
+        return;
+    }
+    if (ImGui::Button("Refresh")) {
+        std::string root = g_app_dir + "/collectors";
+        collectors_discover(root, g_collectors);
+        g_coll_selected = -1;
+    }
+    ImGui::SameLine();
+    ImGui::TextDisabled("%zu collectors", g_collectors.size());
+
+    if (g_collectors.empty()) {
+        ImGui::TextDisabled("No collectors found in %s/collectors/",
+                            g_app_dir.c_str());
+        ImGui::End();
+        return;
+    }
+
+    if (ImGui::BeginTable("##collectors", 3,
+            ImGuiTableFlags_RowBg | ImGuiTableFlags_Borders |
+            ImGuiTableFlags_Resizable)) {
+        ImGui::TableSetupColumn("ID", ImGuiTableColumnFlags_WidthFixed, 180);
+        ImGui::TableSetupColumn("Name", ImGuiTableColumnFlags_WidthFixed, 200);
+        ImGui::TableSetupColumn("Description", ImGuiTableColumnFlags_WidthStretch);
+        ImGui::TableHeadersRow();
+
+        for (int i = 0; i < (int)g_collectors.size(); ++i) {
+            const auto& c = g_collectors[i];
+            ImGui::TableNextRow();
+            ImGui::TableSetColumnIndex(0);
+            bool sel = (i == g_coll_selected);
+            if (ImGui::Selectable(c.id.c_str(), sel,
+                                  ImGuiSelectableFlags_SpanAllColumns)) {
+                g_coll_selected = i;
+            }
+            ImGui::TableSetColumnIndex(1);
+            ImGui::TextUnformatted(c.name.c_str());
+            ImGui::TableSetColumnIndex(2);
+            ImGui::TextUnformatted(c.description.c_str());
+        }
+        ImGui::EndTable();
+    }
+
+    if (g_coll_selected >= 0 &&
+        g_coll_selected < (int)g_collectors.size()) {
+        const auto& c = g_collectors[g_coll_selected];
+        ImGui::Separator();
+        ImGui::Text("Selected: %s", c.id.c_str());
+        ImGui::TextDisabled("%s", c.run_py.c_str());
+
+        ImGui::PushItemWidth(120);
+        ImGui::InputInt("limit", &g_coll_param_limit);
+        ImGui::PopItemWidth();
+        if (g_coll_param_limit < 1) g_coll_param_limit = 1;
+        if (g_coll_param_limit > 500) g_coll_param_limit = 500;
+
+        ImGui::BeginDisabled(g_running || !c.has_run);
+        if (ImGui::Button(g_running ? "Running..." : (TI_PLAYER_PLAY " Run"))) {
+            run_selected_collector();
+        }
+        ImGui::EndDisabled();
+        if (!c.has_run) {
+            ImGui::SameLine();
+            ImGui::TextDisabled("(missing run.py)");
+        }
+    }
+
+    if (!g_last_run_summary.empty()) {
+        ImGui::Separator();
+        ImGui::TextUnformatted("Last run:");
+        ImGui::BeginChild("##run_summary", ImVec2(0, 200), true);
+        ImGui::TextUnformatted(g_last_run_summary.c_str());
+        ImGui::EndChild();
+    }
+
+    ImGui::End();
+}
+
 static void draw_launcher() {
     if (!g_show_launcher) return;
     if (!ImGui::Begin(TI_SEARCH " Launcher", &g_show_launcher)) {
@@ -131,6 +263,7 @@ static void draw_datasets() {
 static void render() {
     static fn_ui::PanelToggle panels[] = {
         { "Launcher",   nullptr, &g_show_launcher   },
+        { "Collectors", nullptr, &g_show_collectors },
         { "Jobs",       nullptr, &g_show_jobs       },
         { "Datasets",   nullptr, &g_show_datasets   },
     };
@@ -138,6 +271,7 @@ static void render() {
                        sizeof(panels) / sizeof(panels[0]),
                        nullptr);
     draw_launcher();
+    draw_collectors();
     draw_jobs();
     draw_datasets();
 }
@@ -170,6 +304,44 @@ int main(int argc, char** argv) {
         registry_list_recent(g_registry, 50, g_results);
     }
 
+    // Resolver registry_root, app_dir, python_exe.
+    if (const char* root = std::getenv("FN_REGISTRY_ROOT"); root && *root) {
+        g_registry_root = root;
+    } else if (!g_db_path.empty()) {
+        // Asume registry.db esta en raiz del registry.
+        std::filesystem::path p(g_db_path);
+        g_registry_root = p.parent_path().string();
+    }
+
+    if (const char* d = std::getenv("ODR_APP_DIR"); d && *d) {
+        g_app_dir = d;
+    } else if (!g_registry_root.empty()) {
+        g_app_dir = g_registry_root +
+            "/projects/online_data_recopilation/apps/odr_console";
+    }
+    g_ops_db_path = g_app_dir + "/operations.db";
+
+    if (const char* py = std::getenv("FN_PYTHON"); py && *py) {
+        g_python_exe = py;
+    } else if (!g_registry_root.empty()) {
+        g_python_exe = g_registry_root + "/python/.venv/bin/python3";
+    } else {
+        g_python_exe = "python3";
+    }
+
+    // Discover collectors: prefer app_dir/collectors, fallback a
+    // <exe_dir>/assets/collectors/ (Windows distribuible).
+    if (!g_app_dir.empty()) {
+        collectors_discover(g_app_dir + "/collectors", g_collectors);
+    }
+    if (g_collectors.empty()) {
+        const char* assets = fn::asset_dir();
+        if (assets && *assets) {
+            std::string assets_collectors = std::string(assets) + "/collectors";
+            collectors_discover(assets_collectors, g_collectors);
+        }
+    }
+
     fn::AppConfig cfg;
     cfg.title = "odr_console — online data recopilation";
     cfg.width = 1400;

runner.cpp

@@ -0,0 +1,124 @@
+#include "runner.h"
+
+#include <atomic>
+#include <chrono>
+#include <cstdio>
+#include <cstdlib>
+#include <filesystem>
+#include <fstream>
+#include <sstream>
+
+namespace fs = std::filesystem;
+
+namespace odr {
+
+namespace {
+
+long long now_ms() {
+    using namespace std::chrono;
+    return duration_cast<milliseconds>(
+        system_clock::now().time_since_epoch()).count();
+}
+
+std::string make_uid() {
+    static std::atomic<int> ctr{0};
+    long long ts = now_ms();
+    int n = ctr.fetch_add(1, std::memory_order_relaxed);
+    char buf[64];
+    std::snprintf(buf, sizeof(buf), "%lld_%05d", ts, n);
+    return buf;
+}
+
+bool write_file(const fs::path& p, const std::string& bytes) {
+    std::ofstream f(p, std::ios::binary | std::ios::trunc);
+    if (!f.is_open()) return false;
+    f.write(bytes.data(), (std::streamsize)bytes.size());
+    return f.good();
+}
+
+std::string read_file(const fs::path& p) {
+    std::ifstream f(p, std::ios::binary);
+    if (!f.is_open()) return "";
+    std::ostringstream ss;
+    ss << f.rdbuf();
+    return ss.str();
+}
+
+// Quote para shell: encierra en comillas dobles, escapa los " internos.
+// Suficiente para paths del registry — no tocamos shell metacharacters
+// porque no hay input de usuario directo (paths controlados por la app).
+std::string sh_quote(const std::string& s) {
+    std::string out;
+    out.reserve(s.size() + 2);
+    out.push_back('"');
+    for (char c : s) {
+        if (c == '"' || c == '\\' || c == '$' || c == '`') {
+            out.push_back('\\');
+        }
+        out.push_back(c);
+    }
+    out.push_back('"');
+    return out;
+}
+
+} // namespace
+
+RunResult run_collector(const std::string& python_exe,
+                        const std::string& run_py,
+                        const std::string& tmp_dir,
+                        const std::string& ctx_json) {
+    RunResult r;
+
+    std::error_code ec;
+    fs::create_directories(tmp_dir, ec);
+    if (ec) {
+        r.stderr_str = "create_directories failed: " + ec.message();
+        return r;
+    }
+
+    std::string uid = make_uid();
+    fs::path ctx_path = fs::path(tmp_dir) / ("ctx_" + uid + ".json");
+    fs::path out_path = fs::path(tmp_dir) / ("out_" + uid + ".json");
+    fs::path err_path = fs::path(tmp_dir) / ("err_" + uid + ".log");
+
+    if (!write_file(ctx_path, ctx_json)) {
+        r.stderr_str = "write ctx failed";
+        return r;
+    }
+
+    std::string cmd =
+        sh_quote(python_exe) + " " + sh_quote(run_py)
+      + " < " + sh_quote(ctx_path.string())
+      + " > " + sh_quote(out_path.string())
+      + " 2> " + sh_quote(err_path.string());
+
+    long long t0 = now_ms();
+    int rc = std::system(cmd.c_str());
+    r.duration_ms = now_ms() - t0;
+
+#if defined(__unix__) || defined(__APPLE__)
+    // POSIX: WEXITSTATUS para extraer exit code real.
+    if (rc != -1 && WIFEXITED(rc)) {
+        r.exit_code = WEXITSTATUS(rc);
+    } else {
+        r.exit_code = rc;
+    }
+#else
+    r.exit_code = rc;
+#endif
+
+    r.stdout_str = read_file(out_path);
+    r.stderr_str = read_file(err_path);
+
+    // Cleanup salvo si fallo (dejar para debug).
+    if (r.exit_code == 0) {
+        std::error_code rmec;
+        fs::remove(ctx_path, rmec);
+        fs::remove(out_path, rmec);
+        fs::remove(err_path, rmec);
+    }
+
+    return r;
+}
+
+} // namespace odr

runner.h

@@ -0,0 +1,32 @@
+#pragma once
+
+#include <string>
+
+namespace odr {
+
+struct RunResult {
+    int         exit_code   = -1;
+    long long   duration_ms = 0;
+    std::string stdout_str;
+    std::string stderr_str;
+};
+
+// Synchronous subprocess runner. Bloquea el hilo caller hasta que el
+// subprocess termina. UI debe llamarlo desde un thread propio o aceptar
+// el bloqueo (MVP: aceptar bloqueo).
+//
+// Layout temp:
+//   <tmp_dir>/ctx_<uid>.json   — escrito por runner
+//   <tmp_dir>/out_<uid>.json   — stdout capturado
+//   <tmp_dir>/err_<uid>.log    — stderr capturado
+//
+// Usa std::system. python_exe puede ser path absoluto o ejecutable en PATH.
+// run_py es el path absoluto al run.py del collector.
+//
+// Retorna RunResult con exit_code=0 en exito.
+RunResult run_collector(const std::string& python_exe,
+                        const std::string& run_py,
+                        const std::string& tmp_dir,
+                        const std::string& ctx_json);
+
+} // namespace odr