Added trained model on known files

bin/send_dir.sh

@@ -0,0 +1,118 @@
+#!/usr/bin/env bash
+#
+# send_dir.sh — upload every file in a directory to a running FileServer.
+#
+# Each file is POSTed to /upload as multipart/form-data. By default one file per
+# request (so you get per-file feedback); use -b to batch all files into a
+# single request instead.
+#
+# Usage:
+#   bin/send_dir.sh [-u URL] [-b] [-r] DIR
+#
+#   DIR         directory whose files are uploaded
+#   -u URL      server base URL (default: http://127.0.0.1:8080, or $FS_URL)
+#   -b          batch: send all files in one multipart request
+#   -r          recurse into subdirectories
+#   -j N        upload up to N files concurrently (default: 1). A slow big
+#               file no longer blocks the others; results are tallied at the end.
+#
+# Examples:
+#   bin/send_dir.sh data/samples
+#   bin/send_dir.sh -u http://host:9000 -b -r ~/Downloads
+set -euo pipefail
+
+url="${FS_URL:-http://127.0.0.1:8080}"
+batch=0
+recurse=0
+jobs=1
+
+while getopts ":u:brj:h" opt; do
+  case "$opt" in
+    u) url="$OPTARG" ;;
+    b) batch=1 ;;
+    r) recurse=1 ;;
+    j) jobs="$OPTARG" ;;
+    h) sed -n '2,22p' "$0"; exit 0 ;;
+    \?) echo "unknown option: -$OPTARG" >&2; exit 2 ;;
+    :) echo "option -$OPTARG needs an argument" >&2; exit 2 ;;
+  esac
+done
+shift $((OPTIND - 1))
+
+if ! [[ "$jobs" =~ ^[0-9]+$ ]] || [[ "$jobs" -lt 1 ]]; then
+  echo "-j needs a positive integer, got: $jobs" >&2
+  exit 2
+fi
+
+dir="${1:-}"
+if [[ -z "$dir" ]]; then
+  echo "usage: $0 [-u URL] [-b] [-r] DIR" >&2
+  exit 2
+fi
+if [[ ! -d "$dir" ]]; then
+  echo "not a directory: $dir" >&2
+  exit 2
+fi
+
+endpoint="${url%/}/upload"
+
+# Collect regular files (optionally recursive), NUL-safe for odd filenames.
+mapfile -d '' files < <(
+  if [[ "$recurse" -eq 1 ]]; then
+    find "$dir" -type f -print0
+  else
+    find "$dir" -maxdepth 1 -type f -print0
+  fi
+)
+
+if [[ "${#files[@]}" -eq 0 ]]; then
+  echo "no files found in $dir" >&2
+  exit 1
+fi
+
+echo "uploading ${#files[@]} file(s) from '$dir' to $endpoint"
+
+if [[ "$batch" -eq 1 ]]; then
+  # One request with every file. Each -F field name is arbitrary; the server
+  # keys off the filename, so we just number them.
+  args=()
+  i=0
+  for f in "${files[@]}"; do
+    args+=(-F "f${i}=@${f}")
+    i=$((i + 1))
+  done
+  code=$(curl -sS -o /dev/stderr -w '%{http_code}' "${args[@]}" "$endpoint")
+  echo
+  echo "-> HTTP $code"
+  [[ "$code" == "202" ]]
+else
+  # One request per file, up to $jobs in flight at once. Each upload runs in the
+  # background so a slow (big) file doesn't stall the ones behind it. Every job
+  # writes its HTTP status to a per-file line in $results; we tally after all
+  # jobs drain.
+  results="$(mktemp)"
+  trap 'rm -f "$results"' EXIT
+
+  # Upload one file: print progress and append "code<TAB>path" to $results.
+  upload_one() {
+    local f="$1" code
+    code=$(curl -sS -o /dev/stderr -w '%{http_code}' -F "file=@${f}" "$endpoint")
+    echo "  -> HTTP $code  ${f}"
+    printf '%s\t%s\n' "$code" "$f" >> "$results"
+  }
+
+  for f in "${files[@]}"; do
+    # Throttle: wait for a slot to free up before launching the next upload.
+    # (|| true: a failed upload's exit status must not trip `set -e` here.)
+    while (( $(jobs -r -p | wc -l) >= jobs )); do
+      wait -n || true
+    done
+    upload_one "$f" &
+  done
+  wait || true
+
+  ok=$(grep -c $'^202\t' "$results" || true)
+  fail=$(( ${#files[@]} - ok ))
+  echo "done: $ok accepted, $fail failed"
+  [[ "$fail" -eq 0 ]]
+fi