class

Analyzer::Elixir::Phoenix

Inherits Analyzer::Elixir::ElixirEngine < FileScanEngine < Analyzer < FileHelper < Reference < Object

Constants

ELIXIR_INTERPOLATION_RE = /\#\{([^}]+)\}/

Elixir double-quoted strings interpolate #{expr}, and the Phoenix Router DSL captures the literal characters between quotes. Without normalization, get "/api/#{@version}/items" produced URL /api/#{@version}/items with the #{@…} syntax leaking into the path. Rewrite each interpolation site to a {name} placeholder (stripping any leading @ module- attribute sigil) so the path-parameter extractor recognises it and the URL template reads cleanly. Mirrors the Python f-string, Ruby #{}, PHP $var, and Crystal #{} fixes.

MODULE_NAME_RE = /^\s*defmodule\s+([A-Za-z_]\w*(?:\.[A-Za-z_]\w*)*)\s+do\b/m
PHOENIX_ROUTE_EVIDENCE_RE = /\ \b(?:get|post|put|patch|delete|options|head|match|resources|scope|socket|live|live_dashboard|forward) \s*(?:\(\s*)?(?:["']|unquote\b|:) |\bdefmacro\s+ /x

File-level gate for the route-extraction pass. Controllers, schemas, views, and most application modules never declare Phoenix routes, but analyze_file used to line-scan every .ex file anyway. Match the DSL shape (verb + optional paren + string/unquote/atom) so common non-route uses like Map.get( or forget do not force a full scan. Custom route macros collected in the first pass are checked by name separately in may_contain_phoenix_routes?.

STANDARD_ROUTE_METHODS = {"get" => "GET", "post" => "POST", "patch" => "PATCH", "put" => "PUT", "delete" => "DELETE", "options" => "OPTIONS", "head" => "HEAD"}

get/post/… → its precompiled route regex. add_standard_route runs once per verb per line of every .ex file; Regex.new there recompiled the same 7 PCRE2 patterns on every call (~3.4M recompiles on a 2400-file repo, the dominant scan cost). The pattern depends only on the fixed verb set, so build it once. See add_standard_route for the shape rationale. RFC 10008's query is deliberately absent here. Unlike Plug (which already ships query/2), Phoenix's own router macro is still pending (phoenixframework/phoenix#6737); adding a bare verb here — via add_standard_route, which has no plug_route_path?-style leading-/ guard — would misread any unrelated query("literal", CapitalizedMod) local function (a common Ecto/dispatch-helper shape) as a route. Since a Phoenix router sits on Plug, match :query, ... already routes today without needing this table — see match_route_methods below, which upcases whatever verb atom it's given.

STANDARD_ROUTE_PATTERNS = STANDARD_ROUTE_METHODS.keys.to_h do |verb| {verb, Regex.new("(?:^|[^.\\w])#{verb}\\s*(?:\\(\\s*)?['\"]([^'\"]+)['\"]\\s*,\\s*([A-Z]\\w*(?:\\.[A-Za-z_]\\w*)*|unquote\\(\\s*\\w+\\s*\\))(?=\\s*[,)]|\\s*$)(?:\\s*,\\s*:(\\w+[!?]?))?")} end

Class methods

tech_name
Source

Instance methods

analyze
Source
analyze_file(path : String) : Array(Endpoint)
Source
extract_controller_params
Source
extract_params_from_controller(content : String, controller_name : String, controller_path : String)
Source
extract_params_from_function_block(lines : Array(String), start_index : Int32, end_index : Int32, method : String) : Array(Param)
Source
find_function_end(lines : Array(String), start_index : Int32) : Int32
Source
line_to_endpoint(line : String, file_path : String, scope_prefix : String = "", scope_module : String = "", base : String = @base_path, string_bindings : Hash(String, String) = Hash(String, String).new) : Array(Endpoint)
Source
tech

Instance-side view of the same declaration. The per-file rescues live on this base class, which has no way to name the analyzer that is running inside them, so a skipped file could not be attributed to a tech. Deriving it from analyzer_for keeps the name written exactly once.

Source

Nested types