Merge pull request #189 from barrett-ruth/feat/multi-test-case

Multi-Test Case View

.pre-commit-config.yaml

@@ -2,7 +2,7 @@ minimum_pre_commit_version: '3.5.0'
 
 repos:
   - repo: https://github.com/JohnnyMorganz/StyLua
-    rev: v2.1.0
+    rev: v2.3.1
     hooks:
       - id: stylua-github
         name: stylua (Lua formatter)
@@ -10,7 +10,7 @@ repos:
         pass_filenames: true
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.6.9
+    rev: v0.14.3
     hooks:
       - id: ruff-format
         name: ruff (format)
@@ -30,7 +30,7 @@ repos:
         pass_filenames: false
 
   - repo: https://github.com/pre-commit/mirrors-prettier
-    rev: v3.1.0
+    rev: v4.0.0-alpha.8
     hooks:
       - id: prettier
         name: prettier (format markdown)

doc/cp.nvim.txt

@@ -34,15 +34,30 @@ COMMANDS                                                         *cp-commands*
                                     :CP codeforces 1933 --lang python
 <
         View Commands ~
-            :CP run [--debug] [n]
+            :CP run [all|n|n,m,...] [--debug]
                                 Run tests in I/O view (see |cp-io-view|).
                                 Lightweight split showing test verdicts.
-                                Without [n]: runs all tests, shows verdict summary
+
-                                With [n]: runs test n, shows detailed output
+                                Execution modes:
+                                • :CP run              Combined: single execution with all tests
+                                                       (auto-switches to individual when multiple samples)
+                                • :CP run all          Individual: N separate executions
+                                • :CP run n            Individual: run test n only
+                                • :CP run n,m,...      Individual: run specific tests (e.g. nth and mth)
+
                                 --debug: Use debug build (builds to build/<name>.dbg)
+
+                                Combined mode runs all test inputs in one execution (matching
+                                platform behavior for multi-test problems). When a problem has
+                                multiple independent sample test cases, :CP run auto-switches to
+                                individual mode to run each sample separately.
+
                                 Examples: >
-                                    :CP run              " All tests
+                                    :CP run              " Combined: all tests, one execution
-                                    :CP run --debug 2    " Test 2, debug build
+                                    :CP run all          " Individual: all tests, N executions
+                                    :CP run 2            " Individual: test 2 only
+                                    :CP run 1,3,5        " Individual: tests 1, 3, and 5
+                                    :CP run all --debug  " Individual with debug build
 <
             :CP panel [--debug] [n]
                                 Open full-screen test panel (see |cp-panel|).
@@ -536,10 +551,27 @@ Example: Setting up and solving AtCoder contest ABC324
 I/O VIEW                                                          *cp-io-view*
 
 The I/O view provides lightweight test feedback in persistent side splits.
-All test outputs are concatenated with verdict summaries at the bottom.
+Test outputs are concatenated with verdict summaries at the bottom.
 The |cp-panel| offers more fine-grained analysis with diff modes.
 
-Access the I/O view with :CP run [n]
+Execution Modes ~
+
+The I/O view supports two execution modes:
+
+Combined Mode (:CP run with single sample)
+    • Single execution with all test inputs concatenated
+    • Matches platform behavior (e.g. Codeforces multi-test format)
+    • Shows one verdict for the entire execution
+    • Input split: All test inputs concatenated
+    • Output split: Single program output + verdict
+    • Used when problem has one sample containing multiple test cases
+
+Individual Mode (:CP run all / :CP run n / :CP run n,m,...)
+    • Separate execution for each test case
+    • Per-test verdicts for debugging
+    • Input split: Selected test inputs concatenated
+    • Output split: All test outputs concatenated + per-test verdicts
+    • Auto-selected when problem has multiple independent samples
 
 Layout ~
 
@@ -561,7 +593,7 @@ The I/O view appears as 30% width splits on the right side: >
     └──────────────────────────┴─────────────────────────────────────────────┘
 <
 The output split shows:
-1. Concatenated test outputs (separated by blank lines)
+1. Program output (raw, preserving all formatting)
 2. Space-aligned verdict summary with:
    - Test number and status (AC/WA/TLE/MLE/RTE with color highlighting)
    - Runtime: actual/limit in milliseconds
@@ -570,8 +602,10 @@ The output split shows:
 
 Usage ~
 
-    :CP run             Run all tests
+    :CP run             Combined mode: all tests in one execution
-    :CP run 3           Run test 3 only
+    :CP run all         Individual mode: all tests separately
+    :CP run 3           Individual mode: test 3 only
+    :CP run 1,3,5       Individual mode: specific tests (1, 3, and 5)
 
 Navigation ~
 

lua/cp/cache.lua

@@ -16,6 +16,10 @@
 ---@field name string
 ---@field id string
 
+---@class CombinedTest
+---@field input string
+---@field expected string
+
 ---@class Problem
 ---@field id string
 ---@field name? string
@@ -23,6 +27,7 @@
 ---@field multi_test? boolean
 ---@field memory_mb? number
 ---@field timeout_ms? number
+---@field combined_test? CombinedTest
 ---@field test_cases TestCase[]
 
 ---@class TestCase
@@ -181,9 +186,34 @@ function M.get_test_cases(platform, contest_id, problem_id)
   return cache_data[platform][contest_id].problems[index].test_cases or {}
 end
 
+---@param platform string
+---@param contest_id string
+---@param problem_id? string
+---@return CombinedTest?
+function M.get_combined_test(platform, contest_id, problem_id)
+  vim.validate({
+    platform = { platform, 'string' },
+    contest_id = { contest_id, 'string' },
+    problem_id = { problem_id, { 'string', 'nil' }, true },
+  })
+
+  if
+    not cache_data[platform]
+    or not cache_data[platform][contest_id]
+    or not cache_data[platform][contest_id].problems
+    or not cache_data[platform][contest_id].index_map
+  then
+    return nil
+  end
+
+  local index = cache_data[platform][contest_id].index_map[problem_id]
+  return cache_data[platform][contest_id].problems[index].combined_test
+end
+
 ---@param platform string
 ---@param contest_id string
 ---@param problem_id string
+---@param combined_test? CombinedTest
 ---@param test_cases TestCase[]
 ---@param timeout_ms number
 ---@param memory_mb number
@@ -193,6 +223,7 @@ function M.set_test_cases(
   platform,
   contest_id,
   problem_id,
+  combined_test,
   test_cases,
   timeout_ms,
   memory_mb,
@@ -203,6 +234,7 @@ function M.set_test_cases(
     platform = { platform, 'string' },
     contest_id = { contest_id, 'string' },
     problem_id = { problem_id, { 'string', 'nil' }, true },
+    combined_test = { combined_test, { 'table', 'nil' }, true },
     test_cases = { test_cases, 'table' },
     timeout_ms = { timeout_ms, { 'number', 'nil' }, true },
     memory_mb = { memory_mb, { 'number', 'nil' }, true },
@@ -212,6 +244,7 @@ function M.set_test_cases(
 
   local index = cache_data[platform][contest_id].index_map[problem_id]
 
+  cache_data[platform][contest_id].problems[index].combined_test = combined_test
   cache_data[platform][contest_id].problems[index].test_cases = test_cases
   cache_data[platform][contest_id].problems[index].timeout_ms = timeout_ms
   cache_data[platform][contest_id].problems[index].memory_mb = memory_mb

lua/cp/commands/init.lua

@@ -17,8 +17,11 @@ local actions = constants.ACTIONS
 ---@field problem_id? string
 ---@field interactor_cmd? string
 ---@field test_index? integer
+---@field test_indices? integer[]
+---@field mode? string
 ---@field debug? boolean
 ---@field language? string
+---@field subcommand? string
 
 --- Turn raw args into normalized structure to later dispatch
 ---@param args string[] The raw command-line mode args
@@ -75,25 +78,84 @@ local function parse_command(args)
       return { type = 'action', action = 'edit', test_index = test_index }
     elseif first == 'run' or first == 'panel' then
       local debug = false
-      local test_index = nil
+      local test_indices = nil
+      local mode = 'combined'
 
       if #args == 2 then
         if args[2] == '--debug' then
           debug = true
+        elseif args[2] == 'all' then
+          mode = 'individual'
+        else
+          if args[2]:find(',') then
+            local indices = {}
+            for num in args[2]:gmatch('[^,]+') do
+              local idx = tonumber(num)
+              if not idx or idx < 1 or idx ~= math.floor(idx) then
+                return {
+                  type = 'error',
+                  message = ("Invalid test index '%s' in list"):format(num),
+                }
+              end
+              table.insert(indices, idx)
+            end
+            if #indices == 0 then
+              return { type = 'error', message = 'No valid test indices provided' }
+            end
+            test_indices = indices
+            mode = 'individual'
           else
             local idx = tonumber(args[2])
             if not idx then
               return {
                 type = 'error',
-              message = ("Invalid argument '%s': expected test number or --debug"):format(args[2]),
+                message = ("Invalid argument '%s': expected test number(s), 'all', or --debug"):format(
+                  args[2]
+                ),
               }
             end
             if idx < 1 or idx ~= math.floor(idx) then
               return { type = 'error', message = ("'%s' is not a valid test index"):format(idx) }
             end
-          test_index = idx
+            test_indices = { idx }
+            mode = 'individual'
+          end
         end
       elseif #args == 3 then
+        if args[2] == 'all' then
+          mode = 'individual'
+          if args[3] ~= '--debug' then
+            return {
+              type = 'error',
+              message = ("Invalid argument '%s': expected --debug"):format(args[3]),
+            }
+          end
+          debug = true
+        elseif args[2]:find(',') then
+          local indices = {}
+          for num in args[2]:gmatch('[^,]+') do
+            local idx = tonumber(num)
+            if not idx or idx < 1 or idx ~= math.floor(idx) then
+              return {
+                type = 'error',
+                message = ("Invalid test index '%s' in list"):format(num),
+              }
+            end
+            table.insert(indices, idx)
+          end
+          if #indices == 0 then
+            return { type = 'error', message = 'No valid test indices provided' }
+          end
+          if args[3] ~= '--debug' then
+            return {
+              type = 'error',
+              message = ("Invalid argument '%s': expected --debug"):format(args[3]),
+            }
+          end
+          test_indices = indices
+          mode = 'individual'
+          debug = true
+        else
           local idx = tonumber(args[2])
           if not idx then
             return {
@@ -110,16 +172,26 @@ local function parse_command(args)
               message = ("Invalid argument '%s': expected --debug"):format(args[3]),
             }
           end
-        test_index = idx
+          test_indices = { idx }
+          mode = 'individual'
           debug = true
+        end
       elseif #args > 3 then
         return {
           type = 'error',
-          message = 'Too many arguments. Usage: :CP ' .. first .. ' [test_num] [--debug]',
+          message = 'Too many arguments. Usage: :CP '
+            .. first
+            .. ' [all|test_num[,test_num...]] [--debug]',
         }
       end
 
-      return { type = 'action', action = first, test_index = test_index, debug = debug }
+      return {
+        type = 'action',
+        action = first,
+        test_indices = test_indices,
+        debug = debug,
+        mode = mode,
+      }
     else
       local language = nil
       if #args >= 3 and args[2] == '--lang' then
@@ -197,9 +269,12 @@ function M.handle_command(opts)
     if cmd.action == 'interact' then
       ui.toggle_interactive(cmd.interactor_cmd)
     elseif cmd.action == 'run' then
-      ui.run_io_view(cmd.test_index, cmd.debug)
+      ui.run_io_view(cmd.test_indices, cmd.debug, cmd.mode)
     elseif cmd.action == 'panel' then
-      ui.toggle_panel({ debug = cmd.debug, test_index = cmd.test_index })
+      ui.toggle_panel({
+        debug = cmd.debug,
+        test_index = cmd.test_indices and cmd.test_indices[1] or nil,
+      })
     elseif cmd.action == 'next' then
       setup.navigate_problem(1, cmd.language)
     elseif cmd.action == 'prev' then

lua/cp/runner/run.lua

@@ -198,6 +198,40 @@ function M.load_test_cases()
   return #tcs > 0
 end
 
+---@param debug boolean?
+---@return RanTestCase?
+function M.run_combined_test(debug)
+  local combined = cache.get_combined_test(
+    state.get_platform() or '',
+    state.get_contest_id() or '',
+    state.get_problem_id()
+  )
+
+  if not combined then
+    logger.log('No combined test found', vim.log.levels.ERROR)
+    return nil
+  end
+
+  local ran_test = {
+    index = 1,
+    input = combined.input,
+    expected = combined.expected,
+    status = 'running',
+    actual = nil,
+    time_ms = nil,
+    code = nil,
+    ok = nil,
+    signal = nil,
+    tled = false,
+    mled = false,
+    rss_mb = 0,
+    selected = true,
+  }
+
+  local result = run_single_test_case(ran_test, debug)
+  return result
+end
+
 ---@param index number
 ---@param debug boolean?
 ---@return boolean

lua/cp/scraper.lua

@@ -194,6 +194,7 @@ function M.scrape_all_tests(platform, contest_id, callback)
         end
         if type(callback) == 'function' then
           callback({
+            combined = ev.combined,
             tests = ev.tests,
             timeout_ms = ev.timeout_ms or 0,
             memory_mb = ev.memory_mb or 0,

lua/cp/setup.lua

@@ -82,7 +82,7 @@ local function start_tests(platform, contest_id, problems)
     return not vim.tbl_isempty(cache.get_test_cases(platform, contest_id, p.id))
   end, problems)
   if cached_len ~= #problems then
-    logger.log(('Fetching test cases... (%d/%d)'):format(cached_len, #problems))
+    logger.log(('Fetching problem test data... (%d/%d)'):format(cached_len, #problems))
     scraper.scrape_all_tests(platform, contest_id, function(ev)
       local cached_tests = {}
       if not ev.interactive and vim.tbl_isempty(ev.tests) then
@@ -95,6 +95,7 @@ local function start_tests(platform, contest_id, problems)
         platform,
         contest_id,
         ev.problem_id,
+        ev.combined,
         cached_tests,
         ev.timeout_ms or 0,
         ev.memory_mb or 0,
@@ -104,31 +105,12 @@ local function start_tests(platform, contest_id, problems)
 
       local io_state = state.get_io_view_state()
       if io_state then
-        local problem_id = state.get_problem_id()
+        local combined_test = cache.get_combined_test(platform, contest_id, state.get_problem_id())
-        local test_cases = cache.get_test_cases(platform, contest_id, problem_id)
+        if combined_test then
-        local input_lines = {}
+          local input_lines = vim.split(combined_test.input, '\n')
-
-        local contest_data = cache.get_contest_data(platform, contest_id)
-        local is_multi_test = contest_data.problems[contest_data.index_map[problem_id]].multi_test
-
-        if is_multi_test and #test_cases > 1 then
-          table.insert(input_lines, tostring(#test_cases))
-          for _, tc in ipairs(test_cases) do
-            local stripped = tc.input:gsub('^1\n', '')
-            for _, line in ipairs(vim.split(stripped, '\n')) do
-              table.insert(input_lines, line)
-            end
-          end
-        else
-          for _, tc in ipairs(test_cases) do
-            for _, line in ipairs(vim.split(tc.input, '\n')) do
-              table.insert(input_lines, line)
-            end
-          end
-        end
-
           require('cp.utils').update_buffer_content(io_state.input_buf, input_lines, nil, nil)
         end
+      end
     end)
   end
 end

lua/cp/ui/edit.lua

@@ -274,10 +274,25 @@ local function save_all_tests()
   local is_multi_test = contest_data.problems[contest_data.index_map[problem_id]].multi_test
     or false
 
+  -- Generate combined test from individual test cases
+  local combined_input = table.concat(
+    vim.tbl_map(function(tc)
+      return tc.input
+    end, edit_state.test_cases),
+    '\n'
+  )
+  local combined_expected = table.concat(
+    vim.tbl_map(function(tc)
+      return tc.expected
+    end, edit_state.test_cases),
+    '\n'
+  )
+
   cache.set_test_cases(
     platform,
     contest_id,
     problem_id,
+    { input = combined_input, expected = combined_expected },
     edit_state.test_cases,
     edit_state.constraints and edit_state.constraints.timeout_ms or 0,
     edit_state.constraints and edit_state.constraints.memory_mb or 0,

lua/cp/ui/views.lua

@@ -287,7 +287,7 @@ function M.ensure_io_view()
         return
       end
       io_view_state.current_test_index = new_index
-      M.run_io_view(new_index)
+      M.run_io_view({ new_index }, false, 'individual')
     end
 
     if cfg.ui.run.next_test_key then
@@ -338,7 +338,9 @@ function M.ensure_io_view()
   vim.api.nvim_set_current_win(solution_win)
 end
 
-function M.run_io_view(test_index, debug)
+function M.run_io_view(test_indices_arg, debug, mode)
+  mode = mode or 'combined'
+
   local platform, contest_id, problem_id =
     state.get_platform(), state.get_contest_id(), state.get_problem_id()
   if not platform or not contest_id or not problem_id then
@@ -356,35 +358,56 @@ function M.run_io_view(test_index, debug)
     return
   end
 
+  if mode == 'combined' then
+    local test_cases = cache.get_test_cases(platform, contest_id, problem_id)
+    if test_cases and #test_cases > 1 then
+      mode = 'individual'
+    end
+  end
+
   M.ensure_io_view()
 
   local run = require('cp.runner.run')
+
+  if mode == 'combined' then
+    local combined = cache.get_combined_test(platform, contest_id, problem_id)
+    if not combined then
+      logger.log('No combined test available', vim.log.levels.ERROR)
+      return
+    end
+  else
     if not run.load_test_cases() then
       logger.log('No test cases available', vim.log.levels.ERROR)
       return
     end
+  end
 
-  local test_state = run.get_panel_state()
   local test_indices = {}
 
-  if test_index then
+  if mode == 'individual' then
-    if test_index < 1 or test_index > #test_state.test_cases then
+    local test_state = run.get_panel_state()
+
+    if test_indices_arg then
+      for _, idx in ipairs(test_indices_arg) do
+        if idx < 1 or idx > #test_state.test_cases then
           logger.log(
             string.format(
               'Test %d does not exist (only %d tests available)',
-          test_index,
+              idx,
               #test_state.test_cases
             ),
             vim.log.levels.WARN
           )
           return
         end
-    test_indices = { test_index }
+      end
+      test_indices = test_indices_arg
     else
       for i = 1, #test_state.test_cases do
         test_indices[i] = i
       end
     end
+  end
 
   local io_state = state.get_io_view_state()
   if not io_state then
@@ -418,8 +441,6 @@ function M.run_io_view(test_index, debug)
     return
   end
 
-  run.run_all_test_cases(test_indices, debug)
-
   local run_render = require('cp.runner.run_render')
   run_render.setup_highlights()
 
@@ -430,6 +451,72 @@ function M.run_io_view(test_index, debug)
 
   local formatter = config.ui.run.format_verdict
 
+  if mode == 'combined' then
+    local combined = cache.get_combined_test(platform, contest_id, problem_id)
+
+    if not combined then
+      logger.log('No combined test found', vim.log.levels.ERROR)
+      return
+    end
+
+    run.load_test_cases()
+
+    local result = run.run_combined_test(debug)
+
+    if not result then
+      logger.log('Failed to run combined test', vim.log.levels.ERROR)
+      return
+    end
+
+    input_lines = vim.split(combined.input, '\n')
+
+    if result.actual then
+      output_lines = vim.split(result.actual, '\n')
+    end
+
+    local status = run_render.get_status_info(result)
+    local test_state = run.get_panel_state()
+
+    ---@type VerdictFormatData
+    local format_data = {
+      index = 1,
+      status = status,
+      time_ms = result.time_ms or 0,
+      time_limit_ms = test_state.constraints and test_state.constraints.timeout_ms or 0,
+      memory_mb = result.rss_mb or 0,
+      memory_limit_mb = test_state.constraints and test_state.constraints.memory_mb or 0,
+      exit_code = result.code or 0,
+      signal = (result.code and result.code >= 128)
+          and require('cp.constants').signal_codes[result.code]
+        or nil,
+      time_actual_width = #string.format('%.2f', result.time_ms or 0),
+      time_limit_width = #tostring(
+        test_state.constraints and test_state.constraints.timeout_ms or 0
+      ),
+      mem_actual_width = #string.format('%.0f', result.rss_mb or 0),
+      mem_limit_width = #string.format(
+        '%.0f',
+        test_state.constraints and test_state.constraints.memory_mb or 0
+      ),
+    }
+
+    local verdict_result = formatter(format_data)
+    table.insert(verdict_lines, verdict_result.line)
+
+    if verdict_result.highlights then
+      for _, hl in ipairs(verdict_result.highlights) do
+        table.insert(verdict_highlights, {
+          line_offset = #verdict_lines - 1,
+          col_start = hl.col_start,
+          col_end = hl.col_end,
+          group = hl.group,
+        })
+      end
+    end
+  else
+    run.run_all_test_cases(test_indices, debug)
+    local test_state = run.get_panel_state()
+
     local max_time_actual = 0
     local max_time_limit = 0
     local max_mem_actual = 0
@@ -449,21 +536,28 @@ function M.run_io_view(test_index, debug)
       )
     end
 
-  local is_multi_test = contest_data.problems[contest_data.index_map[problem_id]].multi_test
+    local all_outputs = {}
-
-  if is_multi_test and #test_indices > 1 then
-    table.insert(input_lines, tostring(#test_indices))
-  end
-
     for _, idx in ipairs(test_indices) do
       local tc = test_state.test_cases[idx]
 
+      for _, line in ipairs(vim.split(tc.input, '\n')) do
+        table.insert(input_lines, line)
+      end
+
       if tc.actual then
-      for _, line in ipairs(vim.split(tc.actual, '\n', { plain = true, trimempty = false })) do
+        table.insert(all_outputs, tc.actual)
+      end
+    end
+
+    local combined_output = table.concat(all_outputs, '')
+    if combined_output ~= '' then
+      for _, line in ipairs(vim.split(combined_output, '\n')) do
         table.insert(output_lines, line)
       end
     end
 
+    for _, idx in ipairs(test_indices) do
+      local tc = test_state.test_cases[idx]
       local status = run_render.get_status_info(tc)
 
       ---@type VerdictFormatData
@@ -496,13 +590,6 @@ function M.run_io_view(test_index, debug)
           })
         end
       end
-
-    local test_input = tc.input
-    if is_multi_test and #test_indices > 1 then
-      test_input = test_input:gsub('^1\n', '')
-    end
-    for _, line in ipairs(vim.split(test_input, '\n')) do
-      table.insert(input_lines, line)
     end
   end
 

scrapers/atcoder.py

@@ -16,6 +16,7 @@ from urllib3.util.retry import Retry
 
 from .base import BaseScraper
 from .models import (
+    CombinedTest,
     ContestListResult,
     ContestSummary,
     MetadataResult,
@@ -70,7 +71,7 @@ def _retry_after_requests(details):
     on_backoff=_retry_after_requests,
 )
 def _fetch(url: str) -> str:
-    r = _session.get(url, headers=HEADERS, timeout=TIMEOUT_SECONDS)
+    r = _session.get(url, headers=HEADERS, timeout=TIMEOUT_SECONDS, verify=False)
     if r.status_code in RETRY_STATUS:
         raise requests.HTTPError(response=r)
     r.raise_for_status()
@@ -242,7 +243,8 @@ def _to_problem_summaries(rows: list[dict[str, str]]) -> list[ProblemSummary]:
 
 async def _fetch_all_contests_async() -> list[ContestSummary]:
     async with httpx.AsyncClient(
-        limits=httpx.Limits(max_connections=100, max_keepalive_connections=100)
+        limits=httpx.Limits(max_connections=100, max_keepalive_connections=100),
+        verify=False,
     ) as client:
         first_html = await _get_async(client, ARCHIVE_URL)
         last = _parse_last_page(first_html)
@@ -313,16 +315,25 @@ class AtcoderScraper(BaseScraper):
                 return
             data = await asyncio.to_thread(_scrape_problem_page_sync, category_id, slug)
             tests: list[TestCase] = data.get("tests", [])
+
+            combined_input = "\n".join(t.input for t in tests)
+            combined_expected = "\n".join(t.expected for t in tests)
+
             print(
                 json.dumps(
                     {
                         "problem_id": letter,
+                        "combined": {
+                            "input": combined_input,
+                            "expected": combined_expected,
+                        },
                         "tests": [
                             {"input": t.input, "expected": t.expected} for t in tests
                         ],
                         "timeout_ms": data.get("timeout_ms", 0),
                         "memory_mb": data.get("memory_mb", 0),
                         "interactive": bool(data.get("interactive")),
+                        "multi_test": False,
                     }
                 ),
                 flush=True,
@@ -364,6 +375,7 @@ async def main_async() -> int:
                 success=False,
                 error="Usage: atcoder.py tests <contest_id>",
                 problem_id="",
+                combined=CombinedTest(input="", expected=""),
                 tests=[],
                 timeout_ms=0,
                 memory_mb=0,

scrapers/base.py

@@ -34,10 +34,13 @@ class BaseScraper(ABC):
     def _create_tests_error(
         self, error_msg: str, problem_id: str = "", url: str = ""
     ) -> TestsResult:
+        from .models import CombinedTest
+
         return TestsResult(
             success=False,
             error=f"{self.platform_name}: {error_msg}",
             problem_id=problem_id,
+            combined=CombinedTest(input="", expected=""),
             tests=[],
             timeout_ms=0,
             memory_mb=0,

scrapers/codechef.py

@@ -11,6 +11,7 @@ from scrapling.fetchers import StealthyFetcher
 
 from .base import BaseScraper
 from .models import (
+    CombinedTest,
     ContestListResult,
     ContestSummary,
     MetadataResult,
@@ -230,14 +231,22 @@ class CodeChefScraper(BaseScraper):
                         memory_mb = 256.0
                         interactive = False
 
+                    combined_input = "\n".join(t.input for t in tests)
+                    combined_expected = "\n".join(t.expected for t in tests)
+
                     return {
                         "problem_id": problem_code,
+                        "combined": {
+                            "input": combined_input,
+                            "expected": combined_expected,
+                        },
                         "tests": [
                             {"input": t.input, "expected": t.expected} for t in tests
                         ],
                         "timeout_ms": timeout_ms,
                         "memory_mb": memory_mb,
                         "interactive": interactive,
+                        "multi_test": False,
                     }
 
             tasks = [run_one(problem_code) for problem_code in problems.keys()]
@@ -279,6 +288,7 @@ async def main_async() -> int:
                 success=False,
                 error="Usage: codechef.py tests <contest_id>",
                 problem_id="",
+                combined=CombinedTest(input="", expected=""),
                 tests=[],
                 timeout_ms=0,
                 memory_mb=0,

scrapers/codeforces.py

@@ -13,6 +13,7 @@ from scrapling.fetchers import StealthyFetcher
 
 from .base import BaseScraper
 from .models import (
+    CombinedTest,
     ContestListResult,
     ContestSummary,
     MetadataResult,
@@ -126,16 +127,12 @@ def _extract_samples(block: Tag) -> tuple[list[TestCase], bool]:
                 )
                 for k in keys
             ]
-            samples_with_prefix = [
+            return samples, True
-                TestCase(input=f"1\n{tc.input}", expected=tc.expected) for tc in samples
-            ]
-            return samples_with_prefix, True
 
     inputs = [_text_from_pre(p) for p in input_pres]
     outputs = [_text_from_pre(p) for p in output_pres]
     n = min(len(inputs), len(outputs))
-    samples = [TestCase(input=inputs[i], expected=outputs[i]) for i in range(n)]
+    return [TestCase(input=inputs[i], expected=outputs[i]) for i in range(n)], False
-    return samples, False
 
 
 def _is_interactive(block: Tag) -> bool:
@@ -164,18 +161,35 @@ def _parse_all_blocks(html: str) -> list[dict[str, Any]]:
         name = _extract_title(b)[1]
         if not letter:
             continue
-        tests, multi_test = _extract_samples(b)
+        raw_samples, is_grouped = _extract_samples(b)
         timeout_ms, memory_mb = _extract_limits(b)
         interactive = _is_interactive(b)
+
+        if is_grouped and raw_samples:
+            combined_input = f"{len(raw_samples)}\n" + "\n".join(
+                tc.input for tc in raw_samples
+            )
+            combined_expected = "\n".join(tc.expected for tc in raw_samples)
+            individual_tests = [
+                TestCase(input=f"1\n{tc.input}", expected=tc.expected)
+                for tc in raw_samples
+            ]
+        else:
+            combined_input = "\n".join(tc.input for tc in raw_samples)
+            combined_expected = "\n".join(tc.expected for tc in raw_samples)
+            individual_tests = raw_samples
+
         out.append(
             {
                 "letter": letter,
                 "name": name,
-                "tests": tests,
+                "combined_input": combined_input,
+                "combined_expected": combined_expected,
+                "tests": individual_tests,
                 "timeout_ms": timeout_ms,
                 "memory_mb": memory_mb,
                 "interactive": interactive,
-                "multi_test": multi_test,
+                "multi_test": is_grouped,
             }
         )
     return out
@@ -252,6 +266,10 @@ class CodeforcesScraper(BaseScraper):
                 json.dumps(
                     {
                         "problem_id": pid,
+                        "combined": {
+                            "input": b.get("combined_input", ""),
+                            "expected": b.get("combined_expected", ""),
+                        },
                         "tests": [
                             {"input": t.input, "expected": t.expected} for t in tests
                         ],
@@ -298,6 +316,7 @@ async def main_async() -> int:
                 success=False,
                 error="Usage: codeforces.py tests <contest_id>",
                 problem_id="",
+                combined=CombinedTest(input="", expected=""),
                 tests=[],
                 timeout_ms=0,
                 memory_mb=0,

scrapers/cses.py

@@ -10,6 +10,7 @@ import httpx
 
 from .base import BaseScraper
 from .models import (
+    CombinedTest,
     ContestListResult,
     ContestSummary,
     MetadataResult,
@@ -233,14 +234,23 @@ class CSESScraper(BaseScraper):
                     except Exception:
                         tests = []
                         timeout_ms, memory_mb, interactive = 0, 0, False
+
+                    combined_input = "\n".join(t.input for t in tests)
+                    combined_expected = "\n".join(t.expected for t in tests)
+
                     return {
                         "problem_id": pid,
+                        "combined": {
+                            "input": combined_input,
+                            "expected": combined_expected,
+                        },
                         "tests": [
                             {"input": t.input, "expected": t.expected} for t in tests
                         ],
                         "timeout_ms": timeout_ms,
                         "memory_mb": memory_mb,
                         "interactive": interactive,
+                        "multi_test": False,
                     }
 
             tasks = [run_one(p.id) for p in problems]
@@ -282,6 +292,7 @@ async def main_async() -> int:
                 success=False,
                 error="Usage: cses.py tests <category>",
                 problem_id="",
+                combined=CombinedTest(input="", expected=""),
                 tests=[],
                 timeout_ms=0,
                 memory_mb=0,

scrapers/models.py

@@ -8,6 +8,13 @@ class TestCase(BaseModel):
     model_config = ConfigDict(extra="forbid")
 
 
+class CombinedTest(BaseModel):
+    input: str
+    expected: str
+
+    model_config = ConfigDict(extra="forbid")
+
+
 class ProblemSummary(BaseModel):
     id: str
     name: str
@@ -46,6 +53,7 @@ class ContestListResult(ScrapingResult):
 
 class TestsResult(ScrapingResult):
     problem_id: str
+    combined: CombinedTest
     tests: list[TestCase] = Field(default_factory=list)
     timeout_ms: int
     memory_mb: float

tests/test_scrapers.py

@@ -61,6 +61,16 @@ def test_scraper_offline_fixture_matrix(run_scraper_offline, scraper, mode):
                 tr = TestsResult.model_validate(obj)
                 assert tr.problem_id != ""
                 assert isinstance(tr.tests, list)
+                assert hasattr(tr, "combined"), "Missing combined field"
+                assert tr.combined is not None, "combined field is None"
+                assert hasattr(tr.combined, "input"), "combined missing input"
+                assert hasattr(tr.combined, "expected"), "combined missing expected"
+                assert isinstance(tr.combined.input, str), "combined.input not string"
+                assert isinstance(tr.combined.expected, str), (
+                    "combined.expected not string"
+                )
+                assert hasattr(tr, "multi_test"), "Missing multi_test field"
+                assert isinstance(tr.multi_test, bool), "multi_test not boolean"
                 validated_any = True
             else:
                 assert "problem_id" in obj
@@ -68,5 +78,17 @@ def test_scraper_offline_fixture_matrix(run_scraper_offline, scraper, mode):
                 assert (
                     "timeout_ms" in obj and "memory_mb" in obj and "interactive" in obj
                 )
+                assert "combined" in obj, "Missing combined field in raw JSON"
+                assert isinstance(obj["combined"], dict), "combined not a dict"
+                assert "input" in obj["combined"], "combined missing input key"
+                assert "expected" in obj["combined"], "combined missing expected key"
+                assert isinstance(obj["combined"]["input"], str), (
+                    "combined.input not string"
+                )
+                assert isinstance(obj["combined"]["expected"], str), (
+                    "combined.expected not string"
+                )
+                assert "multi_test" in obj, "Missing multi_test field in raw JSON"
+                assert isinstance(obj["multi_test"], bool), "multi_test not boolean"
                 validated_any = True
         assert validated_any, "No valid tests payloads validated"