Prior to this commit, cofiguration consisted of static values and
dynamic variables calculated based on the static values. Changing a
static value meant that those variables had to be recalculated. In a
reactive environment where things change a lot, there was danger that
some parts do not get calculated, or that recalculation leads to
overwriting of other variables. Now, all the dynamic config entries,
such as paths within the project tree are delegated to "accessor"
methods that evaluate their value at the right time.

* R/mix(get_fn_comp,get_fn_extr,get_fn_summ): New functions to
evaluate paths to the comprehensive table, extraction file and the
summary file.

* R/api.R: Corresponding changes in here.

* NAMESPACE: Some updates.

* R/shiny-ui-base.R(rt_input,mz_input): Parameter pref now determines
the label of the control.

* NAMESPACE: Update exports

* R/api.R(create_plots,save_plots,report): User-level functions to
generate and save plots, as well as generate a report containing them.

* R/mix.R: Supporting functions for above.

* R/resources.R: Supporting params for above.

* R/api.R(subset_summary): New function. This will subset the summary
table according to the user specification.

(order_spectra): Updated docs.

* R/resources.R (DEF_ORDER_SPECTRA): Summary table should be ordered
by sets first and then the quality and intensity as criteria.

* R/api.R (extr_data): An unecessary set column was removed from the
m$extr$ms table to avoid possible confusion. Extraction deals with
unique masses, so this column listed only the first set (in comp
table) that contained a particular mass. Later, during the prescreen
stage the set is coming back directly from the comp table.

* R/api.R(prescreen): Split the qa$ms table into three: one to hold
ms1 spectra (ms1_spec), one for ms2 spectra (ms2_spec) and a summary
table (former ftable, now summ). This was done due to mismatch in the
number of entries between MS1 and MS2 spectra and in order to keep
summ table lightweight (no complex objects such as other tables).

* R/resources.R(MS2_SPEC_COLS, MS1_SPEC_COLS, DEF_ORDER_SPECTRA):
Constants COLS define the columns of ms?_spec and DEF_ORDER_SPECTRA
lost set as uninteresting column.

The best name for what used to be called ftable is summary table.

* R/run.R: Deleted. None of the functions therein are needed anymore.

* R/run.R (gen): This is deprecated.

The naming scheme should be consistent. Variables involving ms1 should
start with ms1_ . The same goes for ms2_ .

... and fix the columns in ftab.

* R/api.R(sort_spectra): New function to sort the spectral table.

* R/resources.R(DEF_ORDER_SPECTRA): New parameter defining the default
order of the spectral table.

* NAMESPACE: Update exports.

* R/api.R(extr_data): Files column is now added to results of future
extraction calculations.

* R/api.R(extr_data): tmp is now a list of future values, not futures.

* R/mix.R(read_conf): Use conf_trans function to interpret some of the
  prescreening input parameters.

* R/api.R(new_state): Update because of read_conf.

* (run_in_dir): Add call to prescreen function.

* R/api.R (prescreen): New function. This exported funciton performs a
  range of checks on the spectra in m$extr$ms and produces the initial
  m$out$tab$ftab file table.

* R/mix.R (assess_ms1, assess_ms2): New functions to QA MS1 and MS2
  spetral info in the m$qa$ms table. Called from prescreen function.

* R/resources.R(QA_FLAGS,QA_NUM_REAL,QA_COLS): Definitions of columns
  in the quality control table.

* NAMESPACE: exports

* R/mix.R(conf_trans_pres): New function to validate prescreening
  input.

* R/resources.R (CONF_PRES_NUM, CONF_PRES_TU): New parameters holding
  names of dimensionless and time parameters in prescreening list,
  respectively.

* R/extraction.R(extract): Spec column of rms2full has to wrap spec
  list into another list so that data table does not unwind when
  rbinding.

* R/extraction.R(extract): Move QA variables calculation into
  prescreening part and regroup ms2 table into EICs etc.

* R/mix.R(new_state): Add fn_comp and fn_ftab init. Also, define fn
  for extracted data.

* R/resources.R(FN_FTAB): A standard filetable name.

(FN_FTAB_DEF_OUT): Redefine in terms of FN_FTAB.

(EMPTY_CONF): Expand with new definitions.

* R/api.R(mk_comp_tab): Update.

* R/mix.R(new_state): Add new conf field, m$extr$fn, to hold the name
  of the file where the extracted data is saved.

* R/api.R(extr_data): Save extracted data.

* extraction.R(filt_ms2_fine): Create a data table as the
  output. Complete the ms2 and m1 spectra extration.

* api.R(extr_data): Adapt to changes in extraction.R and generate
  tables of ms1 EICs and ms2 spectra.

* R/extraction.R <various places>: I used to filter precursors and so
  on based on ID, but the problem is a single mz in a file is defined
  by ID and adduct, not just ID. Changed that in many places, but code
  still not fully functional (multiple of shorer length warnings)

* R/extraction.R(extr_ms2): Now collecting MS2 belonging to different
  compounds correctly.

* R/api.R(extr_data): Substitue call to future::future with
  m$future (which may be something else than future, depending on the
  configuration options).

The intention now is not to write individual spectra to different
files, but to memorise all of them in a table. Also, do not store them
in fragile MSnbase objects, rather store them in normal tables.

* R/extraction.R(filt_ms2_by_prcs, filt_ms2_by_prcs_ht,
verif_prec_fine_ht, ): Arguments change. Also adaptation and
restructuring due to them being called by the new driver functions
filt_ms2_fine and extr_ms2.

(filt_ms2_fine,extr_ms3): New driver functions.

(extract): Update the top-level driver function to call extr_ms2.

<all the rest>: Bits and pieces of obsolete code may have gotten an
update although it will be removed later.

* R/mix.R(new_state): Add debug entry to be used, for example, in the
  concurrency call.