coverage.json contains a report object, which is a hash where keys are file names (absolute paths), and values are coverage data for that file (the result of json.stringify(collector.fileCoverageFor(filename))) Each entry consists of:

• path - The path to the file. This is an absolute path, and should be the same as the key in the report object.

• s - Hash of statement counts, where keys as statement IDs.

• b - Hash of branch counts, where keys are branch IDs and values are arrays of counts. For an if statement, the value would have two counts; one for the if, and one for the else. Switch statements would have an array of values for each case.

• f - Hash of function counts, where keys are function IDs.

• fnMap - Hash of functions where keys are function IDs, and values are {name, line, loc, skip}, where name is the name of the function, line is the line the function is declared on, and loc is the Location of the function declaration (just the declaration, not the entire function body - see 'Location Objects' below.) If skip is present and true, then this indicates that this function was ignored by a ### instabul ignore ... ### pragma. Note that if a function is not ignored the skip field will be missing entirely.

• statementMap - Hash where keys are statement IDs, and values are Location objects for each statement. The Location for a function definition is really an assignment, and should include the entire function. In addition to the normal location object fields, a statementMap entry can also have an optional skip field.

• branchMap - Hash where keys are branch IDs, and values are {line, type, locations} objects. line is the line the branch starts on. type is the type of the branch (e.g. "if", "switch"). locations is an array of Location objects, one for each possible outcome of the branch. Note for an if statement where there is no else clause, there will still be two locations generated. Istanbul does not generate coverage for the default case of a switch statement if default is not explicitly present in the source code.

• l - Hash of line counts, where keys are the line number.

  locations for an if statement are always 0-length and located at the start of the if (even the location for the "else"). For a switch statement, locations start at the start of the case statement and go to the end of the line before the next case statement (note Istanbul does nothing clever here if a case is missing a break.) Each location in locations can also optionally have a skip: true field to indicate that this branch was ignored.

IDs used in the fnMap, statementMap, and branchMap are sequential integers, starting at 1.

Location objects are a {start: {line, column}, end: {line, column}} object that describes the start and end of a piece of code. Note that line is 1-based, but column is 0-based.