cowrite 0.2.0 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '0802510583d99ef115f4b8b21bcbf2ecb6d9924f0b8da3e5e494c47c74e7f5f3'
-  data.tar.gz: 591ba65a82edf2263336d8fe1cc1114c0462eac8138d0d587160187e6ea72b52
+  metadata.gz: 2cffeb7bd5c44317c2084f80ef1e3e0c9c9356f203bce74b3d4538df3f546fc5
+  data.tar.gz: 8cbf2cebb9b34e5e4e9768f205a4d67714cdd256be12ba7443bb97515a9839fe
 SHA512:
-  metadata.gz: 7c5af9c56e2198b2976af8137f984e99632d52978b1002639e5ab8c6e23e7094b8acef41865d75e04a6802d4463eda1cf041790b0ce91c2c890b47f5c7784c76
-  data.tar.gz: 32aa39788c1ea19c2acb3eb0af9ea836d2d0a54688eac498a915248b2b2446f6ee645cd0ed582572488190adf0dfd3eff53ad35a031906fc379ecda3a92ec360
+  metadata.gz: e723beeae2cbd86dbd381a8ebb7747226b68efec1fac667b2ec0e1455da35549a46007f4b3ecd422bf297b656d5f5dbc5ee2cba5c3b0cc31484732ecdefaf77d
+  data.tar.gz: 19f32448a142bcb65b413168314eceeb0231dfb5db6705a8ee350dd7918a0406035ea5b49ddb2977e73c667827a3420ec77a33026bf68949bc00e75c140aa6d3

data/lib/cowrite/cli.rb

@@ -3,6 +3,7 @@
 require "shellwords"
 require "parallel"
 require "tempfile"
+require "optparse"
 
 class Cowrite
   class CLI
@@ -19,38 +20,73 @@ class Cowrite
     end
 
     def run(argv)
-      argv, files = parse_argv(argv)
-
-      abort "Use only first argument for prompt" if argv.size != 1
-      prompt = argv[0]
+      prompt, files, options = parse_argv(argv)
 
       files = find_files prompt if files.empty?
 
-      # prompting on main thread so we can go 1-by-1
-      finish = lambda do |file, i, diff|
-        # ask user if diff is fine (TODO: add a "no" option and re-prompt somehow)
-        prompt "Diff for #{file}:\n#{color_diff(diff)}Apply diff to #{file}?", ["yes"]
+      if (parallel = options[:parallel])
+        prompt_for_each_file_in_parallel prompt, files, parallel
+      else
+        prompt_with_all_files_in_context prompt, files
+      end
+    end
+
+    private
 
-        with_content_in_file(diff) do |path|
-          # apply diff (force sus changes, do not make backups)
-          cmd = "patch --posix #{file} < #{path}"
-          out = `#{cmd}`
-          return if $?.success?
+    # send all files at once to have a bigger context
+    # TODO: might break if context gets too big
+    def prompt_with_all_files_in_context(prompt, files)
+      diffs = @cowrite.diffs(prompt, files)
+      diffs.each_with_index do |(file, diff), i|
+        last = (i + 1 == diffs.size)
+        prompt_to_apply_diff file, diff, last:
+      end
+    end
 
-          # give the user a chance to copy the tempfile or modify it
-          warn "Patch failed:\n#{cmd}\n#{out}"
-          prompt "Continue ?", ["yes"] unless i + 1 == files.size
+    # run each file in parallel (faster) or all at once ?
+    # prompting on main thread so we don't get parallel prompts
+    def prompt_for_each_file_in_parallel(prompt, files, parallel)
+      finish = lambda do |file, i, diffs|
+        diffs.each_with_index do |(_, diff), j|
+          last = (i + 1 == files.size && j + 1 == diffs.size)
+          prompt_to_apply_diff file, diff, last:
         end
       end
 
-      # TODO: --parallel instead of env
-      # produce diffs in parallel since it is slow
-      Parallel.each files, finish:, threads: Integer(ENV["PARALLEL"] || "10"), progress: true do |file|
-        @cowrite.diff file, prompt
+      # produce diffs in parallel since multiple files will be slow and can confuse the model
+      Parallel.each files, finish:, threads: parallel, progress: true do |file, _i|
+        @cowrite.diffs prompt, [file]
       end
     end
 
-    private
+    def prompt_to_apply_diff(file, diff, last:)
+      # ask user if diff is fine (TODO: add a "no" option and re-prompt somehow)
+      prompt "Diff for #{file}:\n#{color_diff(diff)}Apply diff to #{file}?", ["yes"]
+      return if apply_diff(file, diff)
+
+      # ask user to continue if diff failed to apply (to give time for manual fixes or abort)
+      prompt "Continue ?", ["yes"] unless last
+    end
+
+    # apply diff (force sus changes, do not make backups)
+    def apply_diff(file, diff)
+      with_content_in_file(diff) do |path|
+        ensure_file_exists(file)
+        cmd = "patch --posix #{file} < #{path}"
+        out = `#{cmd}`
+        return true if $?.success?
+
+        # give the user a chance to copy the tempfile or modify it
+        warn "Patch failed:\n#{cmd}\n#{out}"
+        false
+      end
+    end
+
+    def ensure_file_exists(file)
+      return if File.exist?(file)
+      FileUtils.mkdir_p(File.dirname(file))
+      File.write file, ""
+    end
 
     def with_content_in_file(content)
       Tempfile.create("cowrite-diff") do |f|
@@ -76,7 +112,7 @@ class Cowrite
         files.each_with_index.filter_map { |f, i| f if chosen.include?(i.to_s) } + # index
         chosen.grep_v(/^\d+$/) # path
       missing = chosen.reject { |f| File.exist?(f) }
-      abort "Files #{missing} do not exist" if missing.any?
+      warn "Files #{missing} do not exist, assuming they will be created" if missing.any?
       chosen
     end
 
@@ -147,13 +183,42 @@ class Cowrite
       string.gsub(/\e\[(\d+)(;\d+)*m/, "")
     end
 
-    # allow passing files after --
     def parse_argv(argv)
-      if (dash_index = argv.index("--"))
-        [argv[0...dash_index], argv[dash_index + 1..]]
-      else
-        [argv, []]
-      end
+      # allow passing files after --
+      argv, files =
+        if (dash_index = argv.index("--"))
+          [argv[0...dash_index], argv[dash_index + 1..]]
+        else
+          [argv, []]
+        end
+
+      options = {}
+      OptionParser.new do |opts|
+        opts.banner = <<-BANNER.gsub(/^ {10}/, "")
+          WWTD: Travis simulator - faster + no more waiting for build emails
+
+          Usage:
+              wwtd
+
+          Options:
+        BANNER
+        opts.on("-p", "--parallel [COUNT]", Integer, "Run each file on its own in parallel") do |c|
+          options[:parallel] = c
+        end
+        opts.on("-h", "--help", "Show this.") do
+          puts opts
+          exit
+        end
+        opts.on("-v", "--version", "Show Version") do
+          puts WWTD::VERSION
+          exit
+        end
+      end.parse!(argv)
+
+      abort "Use only first argument for prompt" if argv.size != 1
+      prompt = argv[0]
+
+      [prompt, files, options]
     end
   end
 end

data/lib/cowrite/version.rb

@@ -1,4 +1,4 @@
 # frozen_string_literal: true
 class Cowrite
-  VERSION = "0.2.0"
+  VERSION = "0.3.0"
 end

data/lib/cowrite.rb

@@ -17,12 +17,13 @@ class Cowrite
     abort files unless $?.success?
 
     prompt = <<~MSG
-      Your task is to find a subset of files from a given list of file names.
-      Only reply with the subset of files, newline separated, nothing else.
+      Your task is to find files that need to be written or read to solve an LLM-prompt.
 
-      Given this list of files: #{files.split("\n").inspect}
+      Only reply with a list of files, newline separated, nothing else.
 
-      Which subset of files would be useful for this LLM prompt:
+      List of local files: #{files.split("\n").inspect}
+
+      LLM-prompt:
       ```
       #{prompt}
       ```
@@ -33,38 +34,44 @@ class Cowrite
     without_quotes(answer).split("\n").map(&:strip) # llms like to add extra spaces
   end
 
-  def diff(file, prompt)
+  def diffs(prompt, files)
     # - tried "patch format" but that is often invalid
     # - tied "full fixed content" but that is always missing the fix
     # - need "ONLY" or it adds comments
     # - tried asking for a diff but it's always in the wrong direction or has sublet bugs that make it unusable
-    content = File.read file
     prompt = <<~MSG
       Solve this prompt:
       ```
       #{prompt}
       ```
 
-      By changing the content of the file #{file}:
-      ```
-      #{File.read file}
-      ```
+      By changing the content of these files:
+      #{files.map { |f| "#{f}:\n```#{file_read_or_empty f}```" }.join("\n")}
 
-      Reply with ONLY:
-      - all changed content inside a single ``` block
-      - what range lines from the original content need to be removed as "LINES: <start>-<end>"
+      Reply with ONLY a newline separated list of changes in the format:
+      - what range lines from the original content need to be removed as "FILE: <file> LINES: <start>-<end>"
+      - changed chunk inside a "```code" block
     MSG
+
     puts "prompt:#{prompt}" if ENV["DEBUG"]
     answer = send_to_openai(prompt)
     puts "answer:\n#{answer}" if ENV["DEBUG"]
-    lines = answer.match(/LINES: (\d+)-(\d+)/)
-    answer.sub!(/.*\z/, "") # remove "LINES" line
-    section = without_quotes(answer).chomp # remove trailing newline since it always emits with an extra one
-    generate_diff content, section, from: Integer(lines[1]), to: Integer(lines[2])
+
+    # - also getting leading "- " since sometimes the model prefixes that
+    # - getting any kind of block since sometimes the model uses "```go"
+    changes = answer.scan(/-? ?FILE: (\S+) LINES: (\d+)-(\d+)\n```\S+\n(.*?)\n```/m)
+    changes.map do |file, start, finish, diff|
+      content = file_read_or_empty file
+      [file, generate_diff(content, diff, from: Integer(start), to: Integer(finish))]
+    end
   end
 
   private
 
+  def file_read_or_empty(file)
+    File.exist?(file) ? File.read(file) : ""
+  end
+
   # remove ```foo<content>``` wrapping
   def without_quotes(answer)
     answer.strip.sub(/\A```\S*\n(.*)```\z/m, "\\1")

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: cowrite
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.3.0
 platform: ruby
 authors:
 - Michael Grosser
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-10-24 00:00:00.000000000 Z
+date: 2024-10-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: parallel