From 6559b538e5e05cfa199c15d46ca5bd0edd353974 Mon Sep 17 00:00:00 2001
From: Oleg Sidorov <me@whitebox.io>
Date: Tue, 28 Nov 2023 14:42:58 +0100
Subject: [PATCH] server : backport .srt output format (#1565)

This commit adds a support of .srt format to Whisper server. The code is
effectively backported from examples/main. The output mimetype is set to
application/x-subrip as per https://en.wikipedia.org/wiki/SubRip.

Example usage:

  curl 127.0.0.1:8080/inference \
    -H "Content-Type: multipart/form-data" \
    -F file="@<file-path>" \
    -F temperature="0.2" \
    -F response-format="srt"
---
 examples/server/server.cpp | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/examples/server/server.cpp b/examples/server/server.cpp
index e9d3f88b..2d15d4cc 100644
--- a/examples/server/server.cpp
+++ b/examples/server/server.cpp
@@ -11,6 +11,7 @@
 #include <thread>
 #include <vector>
 #include <cstring>
+#include <sstream>
 
 #if defined(_MSC_VER)
 #pragma warning(disable: 4244 4267) // possible loss of data
@@ -657,6 +658,27 @@ int main(int argc, char ** argv) {
             std::string results = output_str(ctx, params, pcmf32s);
             res.set_content(results.c_str(), "text/html");
         }
+        else if (params.response_format == srt_format)
+        {
+            std::stringstream ss;
+            const int n_segments = whisper_full_n_segments(ctx);
+            for (int i = 0; i < n_segments; ++i) {
+                const char * text = whisper_full_get_segment_text(ctx, i);
+                const int64_t t0 = whisper_full_get_segment_t0(ctx, i);
+                const int64_t t1 = whisper_full_get_segment_t1(ctx, i);
+                std::string speaker = "";
+
+                if (params.diarize && pcmf32s.size() == 2)
+                {
+                    speaker = estimate_diarization_speaker(pcmf32s, t0, t1);
+                }
+
+                ss << i + 1 + params.offset_n << "\n";
+                ss << to_timestamp(t0, true) << " --> " << to_timestamp(t1, true) << "\n";
+                ss << speaker << text << "\n\n";
+            }
+            res.set_content(ss.str(), "application/x-subrip");
+        }
         // TODO add more output formats
         else
         {